RE: setup.. release ready?
From: Robert Collins That shouldn't prevent you using a local mirror though. Does it prevent you using a local mirror? thoughtfulhumm, nooo/thoughtful that would work, the only (minor) complication then would be that folks would need to keep the cygwin cache stuff on their machines as well as the setup app (at the moment, everything is stored on my machine which folks using setup via the share too) which, being a (predominate) Microsoft shop might make folks less likely to want to install cygwin... Question: are there any instructions on how to setup a local mirror? do I have to merge the two setup.ini files I get from the internet mirrors? Do I have to run a webserver? (I do anyway, but...) HOWEVER, I'm not against releasing setup as stands (it has far too many improvements) if a release with a fix for this was OK'd fairly soon after a patch (commandline off switch?) was checked in. Well, a command line 'off' switch will do the job. It will require conscious decision.. yep, I'm happy with that. :) J.
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sat, 2003-04-05 at 19:23, Gary R. Van Sickle wrote: 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. Much better. A few remaining nits. a) The class name, as discussed before. b) RECTCC_H__F4098557_9A48_446d_AF28_2BE45D29F68D__INCLUDED_ Uhm, thats plain ugly. c) (optional) It might be an idea to mention that no virtual methods are permitted in the class, and that no additional parents are allowed, as either would break binary interchangeability with the MS RECT class. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] Bigger Chooser Episode 4: The Nobox Strikes Back
On Sat, 2003-04-05 at 19:41, Gary R. Van Sickle wrote: Adds a note_nobox() variant of the note() function which doesn't pop up a message box. Used in an upcoming patch to eliminate the final, separate, Install Complete message box in favor of a more-traditional Finished page. This duplicates the mbox logic. Please extract the common logic to a utility class (preferred) or static function (will accept this). Cheers, Rob P.S. Thanks for breaking this up, it makes it a LOT easier to review. -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
setup release pending...
Max, I'm still waiting on feedback from you. I don't need a patch for the MD5 checksums immediately, but I do need to know if you are working on one, or if you are objecting in principle but not planning on creating a patch. If the former, lets set a date for having the patch by. If the latter, let me know and I'll release setup.exe. Cheers, Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
Robert Collins wrote: On Sat, 2003-04-05 at 19:23, Gary R. Van Sickle wrote: 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. Much better. A few remaining nits. a) The class name, as discussed before. b) RECTCC_H__F4098557_9A48_446d_AF28_2BE45D29F68D__INCLUDED_ Uhm, thats plain ugly. c) (optional) It might be an idea to mention that no virtual methods are permitted in the class, and that no additional parents are allowed, as either would break binary interchangeability with the MS RECT class. Given that it is derived from RECT, I think RECTPP is a good name. Max.
Re: setup release pending...
Robert Collins wrote: Max, I'm still waiting on feedback from you. I don't need a patch for the MD5 checksums immediately, but I do need to know if you are working on one, or if you are objecting in principle but not planning on creating a patch. If the former, lets set a date for having the patch by. If the latter, let me know and I'll release setup.exe. Sorry, thought I had replied, but I can't find the message in the archives. I will try and get a patch done by the time you have finished with Gary's patches. BTW, are you aware that Pierre's security updates (that you refer to in http://sources.redhat.com/ml/cygwin-apps/2003-04/msg00091.html) are NOT in setup yet? Max.
base-files-mketc.sh
Have we heard from the maintainer of base-files-mketc.sh recently? About two months ago there was a discussion about putting Windows paths in the symlinks (to avoid problems when changing /cygdrive and with user dependent mounts), but the script was never updated. Pierre
RE: setup.. release ready?
On Sat, 5 Apr 2003, John Morrison wrote: From: Robert Collins That shouldn't prevent you using a local mirror though. Does it prevent you using a local mirror? thoughtfulhumm, nooo/thoughtful that would work, the only (minor) complication then would be that folks would need to keep the cygwin cache stuff on their machines as well as the setup app (at the moment, everything is stored on my machine which folks using setup via the share too) which, being a (predominate) Microsoft shop might make folks less likely to want to install cygwin... John, It's named a cache for a reason. They don't *need* to keep it, they could just blow it away after every installation. In fact, there's another command-line option that would be useful for shops that keep local mirrors... Question: are there any instructions on how to setup a local mirror? do I have to merge the two setup.ini files I get from the internet mirrors? Do I have to run a webserver? (I do anyway, but...) Yes, you do have to run a web- (or ftp-, or rsync) server. You will have to create a (or modify an existing) setup.ini, so that it contains the information about all the packages that you have on the local mirror (which will allow you to also provide older or newer packages, as well as more than 2-3 versions of each package). [snip] Hope this helps, Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II
Re: setup release pending...
On Sat, 5 Apr 2003, Max Bowsher wrote: Robert Collins wrote: Max, I'm still waiting on feedback from you. I don't need a patch for the MD5 checksums immediately, but I do need to know if you are working on one, or if you are objecting in principle but not planning on creating a patch. If the former, lets set a date for having the patch by. If the latter, let me know and I'll release setup.exe. Sorry, thought I had replied, but I can't find the message in the archives. I will try and get a patch done by the time you have finished with Gary's patches. BTW, are you aware that Pierre's security updates (that you refer to in http://sources.redhat.com/ml/cygwin-apps/2003-04/msg00091.html) are NOT in setup yet? Max. Max, Reading the message you referred to: It can go in the next release, which currently has (*a list of patches*) slated for it. I read it as the patches aren't necessarily in yet, but will go in before the release... Also, about the MD5 sum complaint, there were actually 2 separate complaints: 1) the MD5 summing takes too long (for setup to start up), and 2) the MD5 summing happens in the main thread, so setup stops redrawing when it's obscured temporarily. IMO, these two issues should be addressed separately. Fixing 2 is as easy as doing MD5 summing in a separate thread, just like I've done with postinstall scripts. I would offer to take a look, but don't think I'll have the time for a patch in the next few weeks, and don't want to hold up the release. As for fixing 1, it's been suggested, I believe, that only the packages that are being installed be MD5-summed (just summarizing here). Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
Gary R. Van Sickle wrote: Robert Collins wrote: On Sat, 2003-04-05 at 19:23, Gary R. Van Sickle wrote: 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. Much better. A few remaining nits. a) The class name, as discussed before. b) RECTCC_H__F4098557_9A48_446d_AF28_2BE45D29F68D__INCLUDED_ Uhm, thats plain ugly. It's a GUID, it's supposed to by ugly. Guaranteed to be unique, not going to collide with anything ever. That's how MS's AppWizard does it, and it seems like a good idea to me. Please Rob, if you're going to pick nits, pick on something that could possibly matter. For me? Please? Gary, this doesn't matter much, but please no GUIDs. I suggest you use _RECTPP_H_, as this follows the most common scheme in setup. Also, the file has DOS line endings. If convenient, please d2u it. Otherwise, mention that the committer must d2u it when you resend. Max.
Re: setup.. release ready?
Robert Collins wrote: On Sat, 2003-04-05 at 07:51, John Morrison wrote: I'm waiting to hear back from Max as to whether he intends to offer a patch (and if so, when to do what) for the local cache MD5 checks, but AFAIK thats the only current delay. I also would like to be able to switch off the local cache md5sums, most of my co-workers do a 'local' install from a shared directory (*very* restrictive firewall at work). That shouldn't prevent you using a local mirror though. Does it prevent you using a local mirror? HOWEVER, I'm not against releasing setup as stands (it has far too many improvements) if a release with a fix for this was OK'd fairly soon after a patch (commandline off switch?) was checked in. Well, a command line 'off' switch will do the job. It will require conscious decision.. yep, I'm happy with that. It can go in the next release, which currently has: * post install logging * Pierre's new security updates ^ No, they haven't been committed yet. Check cvs diff -r2.340 -rHEAD ChangeLog. * New icon sizes * resizable chooser (? if we get the split out patches in the next week or so) Not resizable. Just bigger. And on that timescale, I can do the md5-on-install-not-on-cachescan thing. Max.
RE: base-files-mketc.sh
From: Pierre A. Humblet Have we heard from the maintainer of base-files-mketc.sh recently? Yeah - I'm here ;) About two months ago there was a discussion about putting Windows paths in the symlinks (to avoid problems when changing /cygdrive and with user dependent mounts), but the script was never updated. Sorry, I either missed the conclusion of the thread or we never came to one... J.
RE: base-files-mketc.sh
On Sat, 5 Apr 2003, John Morrison wrote: From: Pierre A. Humblet Have we heard from the maintainer of base-files-mketc.sh recently? Yeah - I'm here ;) About two months ago there was a discussion about putting Windows paths in the symlinks (to avoid problems when changing /cygdrive and with user dependent mounts), but the script was never updated. Sorry, I either missed the conclusion of the thread or we never came to one... J. John, FWIW, I agreed with Pierre (and Max): http://cygwin.com/ml/cygwin-apps/2003-02/msg00033.html. Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II
Re: base-files-mketc.sh
On Sat, Apr 05, 2003 at 02:00:33PM -0500, Igor Pechtchanski wrote: On Sat, 5 Apr 2003, John Morrison wrote: From: Pierre A. Humblet Have we heard from the maintainer of base-files-mketc.sh recently? Yeah - I'm here ;) About two months ago there was a discussion about putting Windows paths in the symlinks (to avoid problems when changing /cygdrive and with user dependent mounts), but the script was never updated. Sorry, I either missed the conclusion of the thread or we never came to one... FWIW, I agreed with Pierre (and Max): http://cygwin.com/ml/cygwin-apps/2003-02/msg00033.html. Yeah, it's unfortunate that we have to use win32 paths but it seems like the only solution in this case. cgf
RE: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
Robert Collins wrote: On Sat, 2003-04-05 at 19:23, Gary R. Van Sickle wrote: 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. Much better. A few remaining nits. a) The class name, as discussed before. b) RECTCC_H__F4098557_9A48_446d_AF28_2BE45D29F68D__INCLUDED_ Uhm, thats plain ugly. It's a GUID, it's supposed to by ugly. Guaranteed to be unique, not going to collide with anything ever. That's how MS's AppWizard does it, and it seems like a good idea to me. Please Rob, if you're going to pick nits, pick on something that could possibly matter. For me? Please? c) (optional) It might be an idea to mention that no virtual methods are permitted in the class, and that no additional parents are allowed, as either would break binary interchangeability with the MS RECT class. Given that it is derived from RECT, I think RECTPP is a good name. I concur. -- Gary R. Van Sickle Brewer. Patriot.
[PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. -- Gary R. Van Sickle Brewer. Patriot. 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. RECTPP.h Description: Binary data
[PATCH] Bigger Chooser Episode 4: The Nobox Strikes Back
Adds a note_nobox() variant of the note() function which doesn't pop up a message box. Used in an upcoming patch to eliminate the final, separate, Install Complete message box in favor of a more-traditional Finished page. -- Gary R. Van Sickle Brewer. Patriot. nobox.diff Description: Binary data 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * msg.cc (note_nobox): New function. * msg.h (note_nobox): New function declaration.
Curious code in Installer::installOneSource
I'm working in this area to implement the MD5-on-install feature. Is there any reason why I should not simplify this: char msg[64]; strcpy (msg, Installing); Progress.SetText1 (msg); log (LOG_PLAIN, String (msg) + + source.Cached ()); To this: char msg[] = Installing; Progress.SetText1 (msg); log (LOG_PLAIN, String (msg) + + source.Cached ()); Or even, this: Progress.SetText1 (Installing); log (LOG_PLAIN, String (Installing ) + source.Cached ()); (There is no other use of the msg variable outside the context shown here.) Max.
Re: setup release pending...
Igor Pechtchanski wrote: On Sat, 5 Apr 2003, Max Bowsher wrote: Robert Collins wrote: Max, I'm still waiting on feedback from you. I don't need a patch for the MD5 checksums immediately, but I do need to know if you are working on one, or if you are objecting in principle but not planning on creating a patch. If the former, lets set a date for having the patch by. If the latter, let me know and I'll release setup.exe. Sorry, thought I had replied, but I can't find the message in the archives. I will try and get a patch done by the time you have finished with Gary's patches. BTW, are you aware that Pierre's security updates (that you refer to in http://sources.redhat.com/ml/cygwin-apps/2003-04/msg00091.html) are NOT in setup yet? Max. Max, Reading the message you referred to: It can go in the next release, which currently has (*a list of patches*) slated for it. I read it as the patches aren't necessarily in yet, but will go in before the release... Sounds reasonable. But discussion seems to have fizzled regarding Pierre's patch, so we must be careful it isn't forgotten. Also, about the MD5 sum complaint, there were actually 2 separate complaints: 1) the MD5 summing takes too long (for setup to start up), and 2) the MD5 summing happens in the main thread, so setup stops redrawing when it's obscured temporarily. IMO, these two issues should be addressed separately. Fixing 2 is as easy as doing MD5 summing in a separate thread, just like I've done with postinstall scripts. I would offer to take a look, but don't think I'll have the time for a patch in the next few weeks, and don't want to hold up the release. As for fixing 1, it's been suggested, I believe, that only the packages that are being installed be MD5-summed (just summarizing here). I'm attacking 1, trying to shift MD5-summing to as-needed rather than at-startup. Max.
Re: setup release pending...
On Sun, 2003-04-06 at 01:09, Max Bowsher wrote: If the former, lets set a date for having the patch by. If the latter, let me know and I'll release setup.exe. Sorry, thought I had replied, but I can't find the message in the archives. I will try and get a patch done by the time you have finished with Gary's patches. BTW, are you aware that Pierre's security updates (that you refer to in http://sources.redhat.com/ml/cygwin-apps/2003-04/msg00091.html) are NOT in setup yet? Yes. I think we have some crossed wires. I'm not slating Pierre's second security update for the 200303 branch. Ditto for Gary's patches. I'm open to putting a patch from you in the 200303 branch, should you wish to make one in the next couple of days. If that is too soon for you, I will release the 200303 branch, and your patch can go into HEAD and subsequently the next release of setup. Cheers, Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: setup.. release ready?
On Sat, 2003-04-05 at 08:37, Max Bowsher wrote: Well, a command line 'off' switch will do the job. It will require conscious decision.. yep, I'm happy with that. It can go in the next release, which currently has: * post install logging * Pierre's new security updates ^ No, they haven't been committed yet. Check cvs diff -r2.340 -rHEAD ChangeLog. I know they haven't. We haven't branched setup-200304 yet either, which (hopefully) the next release will be. * New icon sizes * resizable chooser (? if we get the split out patches in the next week or so) Not resizable. Just bigger. And on that timescale, I can do the md5-on-install-not-on-cachescan thing. Ok, so you are ok then with the current snapshot being released. Cool... Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: Curious code in Installer::installOneSource
On Sat, 5 Apr 2003, Max Bowsher wrote: I'm working in this area to implement the MD5-on-install feature. Is there any reason why I should not simplify this: char msg[64]; strcpy (msg, Installing); Progress.SetText1 (msg); log (LOG_PLAIN, String (msg) + + source.Cached ()); To this: char msg[] = Installing; Progress.SetText1 (msg); log (LOG_PLAIN, String (msg) + + source.Cached ()); Or even, this: Progress.SetText1 (Installing); log (LOG_PLAIN, String (Installing ) + source.Cached ()); (There is no other use of the msg variable outside the context shown here.) Max. Max, Normally I would argue for a #define (using the Redundancy leads to inconsistency principle), but in this case the two uses seem to be completely independent, so I think your last suggestion is ok. Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II
Re: Curious code in Installer::installOneSource
On Sun, 2003-04-06 at 08:38, Max Bowsher wrote: I'm working in this area to implement the MD5-on-install feature. Is there any reason why I should not simplify this: From memory there was either a const correctness issue, or a funny optimiser bug. Lets keep non-related bugfixes to HEAD though. (i.e. patch that in HEAD not 200303). Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: setup release pending...
On Sun, 2003-04-06 at 08:53, Max Bowsher wrote: I read it as the patches aren't necessarily in yet, but will go in before the release... Sounds reasonable. But discussion seems to have fizzled regarding Pierre's patch, so we must be careful it isn't forgotten. Perhaps the pendig patch list should be resurrected? As for fixing 1, it's been suggested, I believe, that only the packages that are being installed be MD5-summed (just summarizing here). I'm attacking 1, trying to shift MD5-summing to as-needed rather than at-startup. Please, just disable the md5 checks via a command line option. There are non-trivial (non-code) issues to make it default to as-needed. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: setup release pending...
On Sun, 2003-04-06 at 09:15, Pierre A. Humblet wrote: At 11:53 PM 4/5/2003 +0100, Max Bowsher wrote: Sounds reasonable. But discussion seems to have fizzled regarding Pierre's patch, so we must be careful it isn't forgotten. I go by what Rob wrote earlier I've fixed the bug in both HEAD and the release branch. We'll save your second patch for the next release. My second patch needed to be regenerated against the current CVS, due to Rob's intervening changes in the ntsec routine. It's ready. Do you want it now or after the coming release? Now, as it's for HEAD only. Thanks, Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
2nd ntsec patches, take two.
2003-04-05 Pierre Humblet [EMAIL PROTECTED] * main.cc (set_default_sec): Set token owner from token user. Always try to set the token primary group to Users or Admins. Index: main.cc === RCS file: /cvs/cygwin-apps/setup/main.cc,v retrieving revision 2.26 diff -u -p -r2.26 main.cc --- main.cc 2 Apr 2003 14:26:27 - 2.26 +++ main.cc 3 Apr 2003 16:58:44 - @@ -220,18 +220,11 @@ private: Setup::SIDWrapper everyOneSID, administratorsSID, usid; Setup::HANDLEWrapper token; bool failed_; - struct GroupInfo { -GroupInfo() : failed_ (false) {} -void get(Setup::HANDLEWrapper token); -bool failed() const {return failed_;} -void fail() { failed_ = true; } -struct { - PSID psid; - char buf[MAX_SID_LEN]; -} gsid; -DWORD size; -bool failed_; - } primaryGroupInfo; + struct { +PSID psid; +char buf[MAX_SID_LEN]; + } osid; + DWORD size; }; void @@ -338,16 +331,6 @@ NTSecurity::setDefaultDACL () } void -NTSecurity::GroupInfo::get(Setup::HANDLEWrapper token) -{ - if (!GetTokenInformation (token.theHANDLE(), TokenPrimaryGroup, gsid, sizeof gsid, size)) -{ - NoteFailedAPI (GetTokenInformation); - fail(); -} -} - -void NTSecurity::setDefaultSecurity () { @@ -355,42 +338,20 @@ NTSecurity::setDefaultSecurity () if (failed()) return; - primaryGroupInfo.get(token); - if (primaryGroupInfo.failed()) -return; - - /* Get the computer name */ - char compname[MAX_COMPUTERNAME_LENGTH + 1]; - DWORD size = sizeof (compname); - if (!GetComputerName (compname, size)) + /* Get the user */ + if (!GetTokenInformation (token.theHANDLE(), TokenUser, osid, + sizeof osid, size)) { - NoteFailedAPI(GetComputerName); + NoteFailedAPI(GetTokenInformation); return; } - - /* Get the local domain SID */ - SID_NAME_USE use; - char domain[MAX_COMPUTERNAME_LENGTH + 1]; - char lsid[MAX_SID_LEN]; - size = sizeof (lsid); - DWORD sz = sizeof (domain); - if (!LookupAccountName (NULL, compname, lsid, size, - domain, sz, use)) + /* Make it the owner */ + if (!SetTokenInformation (token.theHANDLE(), TokenOwner, osid, + sizeof osid)) { - NoteFailedAPI(LookupAccountName); + NoteFailedAPI(SetTokenInformation); return; } - /* Create the None SID from the domain SID. - On NT the last subauthority of a domain is -1 and it is replaced by the RID. - On other systems the RID is appended. */ - sz = *GetSidSubAuthorityCount (lsid); - if (*GetSidSubAuthority (lsid, sz -1) != (DWORD) -1) -*GetSidSubAuthorityCount (lsid) = ++sz; - *GetSidSubAuthority (lsid, sz -1) = DOMAIN_GROUP_RID_USERS; - - /* See if the group is None */ - if (!EqualSid (primaryGroupInfo.gsid.psid, lsid)) -return; SID_IDENTIFIER_AUTHORITY sid_auth; sid_auth = (SID_IDENTIFIER_AUTHORITY) { SECURITY_NT_AUTHORITY };
Re: setup release pending...
Robert Collins wrote: On Sun, 2003-04-06 at 08:53, Max Bowsher wrote: I read it as the patches aren't necessarily in yet, but will go in before the release... Sounds reasonable. But discussion seems to have fizzled regarding Pierre's patch, so we must be careful it isn't forgotten. Perhaps the pendig patch list should be resurrected? Will do. As for fixing 1, it's been suggested, I believe, that only the packages that are being installed be MD5-summed (just summarizing here). I'm attacking 1, trying to shift MD5-summing to as-needed rather than at-startup. Please, just disable the md5 checks via a command line option. There are non-trivial (non-code) issues to make it default to as-needed. OK, command line option for 200303, better solution for HEAD. It's past midnight here, though, so patch will have to wait until tomorrow. Max.
Re: setup release pending...
On Sun, 2003-04-06 at 09:31, Max Bowsher wrote: Robert Collins wrote: Perhaps the pendig patch list should be resurrected? Will do. Thanks. Please, just disable the md5 checks via a command line option. There are non-trivial (non-code) issues to make it default to as-needed. OK, command line option for 200303, better solution for HEAD. It's past midnight here, though, so patch will have to wait until tomorrow. That's fine - I'll look for it tomorrow :}. Cheers, Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: setup release pending...
Pierre A. Humblet wrote: At 11:53 PM 4/5/2003 +0100, Max Bowsher wrote: Sounds reasonable. But discussion seems to have fizzled regarding Pierre's patch, so we must be careful it isn't forgotten. I go by what Rob wrote earlier I've fixed the bug in both HEAD and the release branch. We'll save your second patch for the next release. My second patch needed to be regenerated against the current CVS, due to Rob's intervening changes in the ntsec routine. It's ready. Do you want it now or after the coming release? Pierre P.S.: After the renumbering stuff, can I still simply run cvs update to get the latest version? Yes. Max.
ATTN: maintainers of ...
the following packages: XFree86/XFree86-base cmakeemacs fileutils (coreutils?) gnugoguile lynx mc mutt nano ncftppinfo procps python robots tcshtetex/tetex-bin tin vim (mine...) terminfo ncurses/* readline/* I'm sending this notice because ncurses, one of the Base libraries on which lots of things depend, has been version-bumped. Thus, while no action is required on your part right now, you should be aware that the next release of your package that you build will NOT depend on libncurses6 but will instead depend on libncurses7 -- so you'll need to update your setup.hint when the time comes. Repeat: no action is required until you rebuild your package after having updated ncurses on your development machine. Please make a note of it --Chuck
RE: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 10:02, Max Bowsher wrote: But RECTFOO is a completely different sort of thing to SIDWrapper/HANDLEWrapper. Surely we shouldn't call it something that suggests resemblance? Your point is reasonable. However, the goal of both things is to ease the use of a MS struct, and to do so in a C++ style way. In that *key* respect they are the same. The exact syntax used will (naturally) depend on the adaption taking place. And that adaptation is not in a manner analagous to that happening in e.g. HANDLEWrapper*. xxWrapper are full-blown, honest-to-God classes, with destructors and everything, and are not inherited from structs. Virtual functions could be added, one could go nuts, no problems. Not so with RECTPP. And the name RECTPP certainly brought to your attention that *something* unusual was up, and thus served the very purpose you've claimed on several occaisions it doesn't. *Both SIDWrapper and HANDLEWrapper are in violation of the one class == one .cc/.h pair rule, and furthermore are wholly contained in main.cc. You wouldn't let me get away with that. However, I'm starting to think this discussion could rumble on for all eternity, and its not a matter of vital importance. Please call it to a halt with a maintainer's decree if you get bored with it. Well, I've already said what I'm happy with... RECTWrapper? Even though it is a very different beast than other xxxWrappers? I really doubt any of us would be satisfied with that situation for any length of time. If I wanted it to be a RECTWrapper, I'd have made it a RECTWrapper. I consciously did not do so because I wanted it to be as lightweight as possible. Rob, this is what's frustrating here: 1. You, Max, and myself have now typed enough email to have written a dozen RECT wrappers/adapters/whatever. And there's still no Bigger Chooser in cvs. 2. You don't like my multi-include guard, yet only after two iterations of patches and several emails have you finally stated the only multi-include guard that you would accept. And what you would accept is not the current pratice in the current code. I.e., you are expecting me to be a mind-reader. 2a. What matters in a multi-include guard is not how ugly it may be, but that it prevents multiple-inclusion of the header without colliding with another #define. Nobody ever types it. Nobody ever looks at it. It's a one-off (well, two-off). If you demand SETUP_filename_H, fine, whatever, just please don't make me try to guess both that and that it's a deal-breaker. 3. You don't like RECTPP, but again, only after two patch submissions and many emails have you decided to tell me what name you would accept, a name suggested by Max, a name which he has now thought better of and disagrees with. I also disagree with it, but frankly I see no evidence that my input on this subject matters to you. 4. If there is a problem with RECTPP, it's that it isn't a class, but a sort of freakshow, inherited from a C struct, public data members flapping in the breeze, frankly like nothing I've seen or done before, and certainly not great design from a Design Patterns perspective. But on that matter you've not said word one, but rather reject it on the multi-include protection #define and its *name*! I simply can't make sense of this. 5. RECTPP is like 0.01% of the Bigger Chooser patch. Am I going to have to buy more bandwidth from my ISP to handle all the patch ping-pong when we get to PropSheet.cc? As I've said before, the changes are considerable. There are new methods and data members; are the *names* going to make the cut? How many iterations before I hit on the ones that will? Listen Rob, I didn't come here for a fight. But I also didn't come here to be a mind-reader, or be micro-managed to death, or to spit into the wind. I came to improve Setup. I've done it before. My original Big Chooser uber-patch does it. The patches I'm trying to submit now will eventually do it, assuming they ever make the cut. My frustration stems not from the amount of work involved in doing the changes you request, but from the apparent arbitrariness and mind-reading aspects of them. *Please* tell me how to get this file into cvs Rob. I need the following: 1. A name for the struct. I have seen no alternative better than RECTPP, but will defer to whatever name you choose. Just make sure you're happy with it, and have thoroughly considered the fact that it is not similar to the xxxWrapper classes, and cannot be made to be so. 2. Confirmation that SETUP_FILENAME_H will not be rejected as a multi-include protection #define. 3. Confirmation that with the above changes, the file will be approved. 4. Some indication that we won't have to do this x1000 when we actually get to the meat of the Bigger Chooser. I will also add some comments wrt virtual functions etc as you mentioned, which BTW seems to me to be much more important than either of the other two issues we've been on
RE: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 13:01, Gary R. Van Sickle wrote: And that adaptation is not in a manner analagous to that happening in e.g. HANDLEWrapper*. xxWrapper are full-blown, honest-to-God classes, with destructors and everything, and are not inherited from structs. Virtual functions could be added, one could go nuts, no problems. Not so with RECTPP. And the name RECTPP certainly brought to your attention that *something* unusual was up, and thus served the very purpose you've claimed on several occaisions it doesn't. I never claimed that RECTPP did not bring it to my attention. I said that RECTPP is not evocative of your intended use. *Both SIDWrapper and HANDLEWrapper are in violation of the one class == one .cc/.h pair rule, and furthermore are wholly contained in main.cc. You wouldn't let me get away with that. And I wasn't about to add new files during the troubleshooting of the security crash in the setup-200303 branch. Besides which, you aren't in violation of that guideline, and I never suggested you were. RECTWrapper? Even though it is a very different beast than other xxxWrappers? I really doubt any of us would be satisfied with that situation for any length of time. If I wanted it to be a RECTWrapper, I'd have made it a RECTWrapper. I consciously did not do so because I wanted it to be as lightweight as possible. Rob, this is what's frustrating here: 1. You, Max, and myself have now typed enough email to have written a dozen RECT wrappers/adapters/whatever. And there's still no Bigger Chooser in cvs. And there won't be until we've worked through the review process. 2. You don't like my multi-include guard, yet only after two iterations of patches and several emails have you finally stated the only multi-include guard that you would accept. And what you would accept is not the current pratice in the current code. I.e., you are expecting me to be a mind-reader. In my first review: Lastly CINSTALL is deprecated. We renamed setup from cinstall to setup when we created the new cvs repository. cinstall to setup. Wasn't that clear enough? Apparently not. Telling you that resulted in a GUID 2a. What matters in a multi-include guard is not how ugly it may be, but that it prevents multiple-inclusion of the header without colliding with another #define. Nobody ever types it. Nobody ever looks at it. It's a one-off (well, two-off). If you demand SETUP_filename_H, fine, whatever, just please don't make me try to guess both that and that it's a deal-breaker. I didn't expect you to guess it. I told you that CINSTALL as the leadin was deprecated as we had renamed to SETUP. As for what matters, ugly does matter. 3. You don't like RECTPP, but again, only after two patch submissions and many emails have you decided to tell me what name you would accept, a name suggested by Max, a name which he has now thought better of and disagrees with. I also disagree with it, but frankly I see no evidence that my input on this subject matters to you. In the first review I suggested two names I would accept. In deference to your rejection of both of those I accepted Max's suggestion. You haven't offered ANY alternatives, you've only defended the choice you put forward. I haven't said I won't accept other alternatives, but I have rejected RECTPP. 4. If there is a problem with RECTPP, it's that it isn't a class, but a sort of freakshow, inherited from a C struct, public data members flapping in the breeze, frankly like nothing I've seen or done before, and certainly not great design from a Design Patterns perspective. But on that matter you've not said word one, but rather reject it on the multi-include protection #define and its *name*! I simply can't make sense of this. I queried you on what it was, have accepted it as a valid approach to make other code clean. Crikey, in the very first review you could have followed my requests and had this commited. 5. RECTPP is like 0.01% of the Bigger Chooser patch. Am I going to have to buy more bandwidth from my ISP to handle all the patch ping-pong when we get to PropSheet.cc? As I've said before, the changes are considerable. There are new methods and data members; are the *names* going to make the cut? How many iterations before I hit on the ones that will? Step one: one concept per patch. That will make the review for each part relatively quick a easy. Listen Rob, I didn't come here for a fight. But I also didn't come here to be a mind-reader, or be micro-managed to death, or to spit into the wind. I came to improve Setup. I've done it before. My original Big Chooser uber-patch does it. The patches I'm trying to submit now will eventually do it, assuming they ever make the cut. My frustration stems not from the amount of work involved in doing the changes you request, but from the apparent arbitrariness and mind-reading aspects of them. Gary, I'm getting frustrated too. I'm
RE: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 13:40, Robert Collins wrote: On Sun, 2003-04-06 at 13:01, Gary R. Van Sickle wrote: And that adaptation is not in a manner analagous to that happening in e.g. HANDLEWrapper*. xxWrapper are full-blown, honest-to-God classes, with destructors and everything, and are not inherited from structs. Virtual functions could be added, one could go nuts, no problems. Not so with RECTPP. And the name RECTPP certainly brought to your attention that *something* unusual was up, and thus served the very purpose you've claimed on several occaisions it doesn't. I never claimed that RECTPP did not bring it to my attention. I said that RECTPP is not evocative of your intended use. Oh, just thinking, one name comes to mind that you might like, and that will evoke the right idea: Setup::RECT. I.E. Put the class in namespace Setup. Then inherit from ::RECT. The comment on virtual use etc near the class defn. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 01:01, Max Bowsher wrote: Robert Collins wrote: On Sat, 2003-04-05 at 19:23, Gary R. Van Sickle wrote: 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. Much better. A few remaining nits. a) The class name, as discussed before. Given that it is derived from RECT, I think RECTPP is a good name. I don't. http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00762.html http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00749.html PP doesn't evoke 'Memory compatible wrapper'. It doesn't even evoke Wrapper. Or Adapter. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
RE: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 01:46, Gary R. Van Sickle wrote: It's a GUID, it's supposed to by ugly. Guaranteed to be unique, not going to collide with anything ever. That's how MS's AppWizard does it, and it seems like a good idea to me. Please Rob, if you're going to pick nits, pick on something that could possibly matter. For me? Please? Sure. And this matters. If you don't change it before submitting, I'll have to change it afterwards. Why should I add to my workload when a simple request here should get it right the first time. This matters. SETUP_FILENAME_H is the preferred way. (We're not a library, so the namespace is ours). c) (optional) It might be an idea to mention that no virtual methods are permitted in the class, and that no additional parents are allowed, as either would break binary interchangeability with the MS RECT class. Given that it is derived from RECT, I think RECTPP is a good name. I concur. I've addressed this already, and in my reply to Max. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
Robert Collins wrote: On Sun, 2003-04-06 at 01:01, Max Bowsher wrote: Robert Collins wrote: On Sat, 2003-04-05 at 19:23, Gary R. Van Sickle wrote: 2003-04-05 Gary R. Van Sickle [EMAIL PROTECTED] * RECTPP.h: New file. Much better. A few remaining nits. a) The class name, as discussed before. Given that it is derived from RECT, I think RECTPP is a good name. I don't. http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00762.html http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00749.html PP doesn't evoke 'Memory compatible wrapper'. It doesn't even evoke Wrapper. Or Adapter. RECTPP - a C++ version of RECT. If that's not enough to convince you, then how about: RECTyou suggest something to put here. Maybe RECTWrapper, and make new files (winwrap.h,.cc ?) to house RECTWrapper, and your SIDWrapper and HANDLEWrapper ? Max.
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 04:13, Max Bowsher wrote: Also, the file has DOS line endings. If convenient, please d2u it. Otherwise, mention that the committer must d2u it when you resend. Good catch, and a very important point. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
Robert Collins wrote: On Sun, 2003-04-06 at 01:46, Gary R. Van Sickle wrote: It's a GUID, it's supposed to by ugly. Guaranteed to be unique, not going to collide with anything ever. That's how MS's AppWizard does it, and it seems like a good idea to me. Please Rob, if you're going to pick nits, pick on something that could possibly matter. For me? Please? Sure. And this matters. If you don't change it before submitting, I'll have to change it afterwards. Why should I add to my workload when a simple request here should get it right the first time. This matters. SETUP_FILENAME_H is the preferred way. (We're not a library, so the namespace is ours). Rather sparsely used at the moment. Want me to do a global rename? Max.
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 09:24, Max Bowsher wrote: (We're not a library, so the namespace is ours). Rather sparsely used at the moment. Want me to do a global rename? Sure, if you have the time. Probably obvious, but this doesn't apply to libgetopt++/zlib/bz2lib/librsync. If you have the time to do this, just do it - ChangeLog + headers, all preapproved. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
Robert Collins wrote: On Sun, 2003-04-06 at 09:22, Max Bowsher wrote: PP doesn't evoke 'Memory compatible wrapper'. It doesn't even evoke Wrapper. Or Adapter. RECTPP - a C++ version of RECT. If it was that, then we could change field, give it child classes (which implies virtual destructors ...) We can't. It's *not* a C++ version of RECT. It's a limited function wrapper. Does it need to be that limited? Maybe not, but for now, if I proposed any changes, potentially large amounts of Garys other code would have to change (beyond the search n replace the class renaming will cause), and that would be unreasonable of me. If that's not enough to convince you, then how about: RECTyou suggest something to put here. Maybe RECTWrapper Exactly. Reviewing my suggestion, I don't like it any more. After all, is doesn't wrap - it inherits. Somehow RECTInheriter just doesn't feel quite right. :-) RECTX? (RECT, Extended) RECTPP? (RECT, Partially Polished) :-) If there is a word describing exactly what a RECTPP is, I don't know it. So why not just give it an abstract name? Then we can just say RECTPP is a RECTPP. Or, call it RECTQR (RECT, Quirkily Retrofitted), if you like. Max.
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 09:51, Max Bowsher wrote: Maybe RECTWrapper Exactly. Reviewing my suggestion, I don't like it any more. After all, is doesn't wrap - it inherits. The Adapter pattern (Design Patterns, pg 139) a.k.a. Wrapper uses inheritance. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
Robert Collins wrote: On Sun, 2003-04-06 at 09:51, Max Bowsher wrote: Maybe RECTWrapper Exactly. Reviewing my suggestion, I don't like it any more. After all, is doesn't wrap - it inherits. The Adapter pattern (Design Patterns, pg 139) a.k.a. Wrapper uses inheritance. But RECTFOO is a completely different sort of thing to SIDWrapper/HANDLEWrapper. Surely we shouldn't call it something that suggests resemblance? However, I'm starting to think this discussion could rumble on for all eternity, and its not a matter of vital importance. Please call it to a halt with a maintainer's decree if you get bored with it. Max.
Re: [PATCH] Bigger Chooser Part 3 In Super 3-D: RECTPP
On Sun, 2003-04-06 at 10:02, Max Bowsher wrote: But RECTFOO is a completely different sort of thing to SIDWrapper/HANDLEWrapper. Surely we shouldn't call it something that suggests resemblance? Your point is reasonable. However, the goal of both things is to ease the use of a MS struct, and to do so in a C++ style way. In that *key* respect they are the same. The exact syntax used will (naturally) depend on the adaption taking place. However, I'm starting to think this discussion could rumble on for all eternity, and its not a matter of vital importance. Please call it to a halt with a maintainer's decree if you get bored with it. Well, I've already said what I'm happy with... Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
RE: Transparent terminal window?
Thanks Biju, Simon. Simon, you're right; ATerm builds easily on Cygwin, but it does have the display problem you mentioned. And on my system, with a very recent fresh install of Cygwin, it maxes out the CPU the second you start it. I haven't even tried to get it to display the transparent background though, have you? I will send a note to ATerm's author to tell him about the CPU bug. Thanks -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Simon Dong Sent: Friday, April 04, 2003 11:13 PM To: [EMAIL PROTECTED] Subject: Re: Transparent terminal window? Aterm doens't work properly on Cygwin/XFree86. It has a problem refreshing the screen. Whatever you type in will not be echoed until you move/resize the window. I got ETerm compile on Cygwin with transparent background. I couldn't get the background image to work (because I couldn't get imlib2 compiled with dll support, I think), but I really don't need it. I had to jump through a lot of hoops though, and didn't get the process documented. For what I can remember, I need to compile edb1.0.3 freetype-1.3.1 imlib2-1.0.6 And apply the kde patch to cygwin for shared mem support, which can be found at kde/cygwin's website. Then I can get Eterm built. -Simon - Original Message - From: Biju G C [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Friday, April 04, 2003 5:26 PM Subject: Re: Transparent terminal window? I have not seen any binary avialable on cygwin try to compile http://aterm.sourceforge.net/ for more screenshots http://xplanet.sourceforge.net/ cheers biju CyGNOME = Cygwin + GNOME http://cygnome.sourceforge.net/ --- Jean-Claude Gervais [EMAIL PROTECTED] wrote: Hi, I'm running WindowMaker, and I was wondering; does anyone know how to get transparent terminal windows? Can you do it with stock Cygwin-XFree or do you have to download and compile a specific terminal emulator? Thanks in advance. __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com
Re: keyboard non-US
kaio wrote: I'll add the values to the list. what is xkb keymap? I meant xkb layout. This is the code of the keyboard layout. And this is it for italy. bye ago NP: VNV Nation - Epicentre -- [EMAIL PROTECTED] http://www.gotti.org ICQ: 126018723
Re: keyboard non-US
kaio wrote: KeyboardType: 4: IBM enhanced (101- or 102-key) keyboard KeyboardLayout: 0410 @Harold: can you apply this patch please? -- --- winconfig.olSat Apr 5 16:13:13 2003 +++ winconfig.c Sat Apr 5 16:11:23 2003 @@ -232,6 +232,7 @@ {0x40b, 4, pc105, fi, NULL, NULL, Finnish}, {0x40c, 4, pc105, fr, NULL, NULL, French (Standard)}, {0x80c, 4, pc105, be, NULL, NULL, French (Belgian)}, +{0x410, 4, pc105, it, NULL, NULL, Italian}, {0x41d, 4, pc105, se, NULL, NULL, Swedish (Sweden)}, {-1, -1, NULL, NULL, NULL, NULL, NULL} }; -- bye ago NP: VNV Nation - Epicentre -- [EMAIL PROTECTED] http://www.gotti.org ICQ: 126018723
Re: keyboard non-US
winconfig.c: where? can you apply this patch please? -- --- winconfig.olSat Apr 5 16:13:13 2003 +++ winconfig.c Sat Apr 5 16:11:23 2003 @@ -232,6 +232,7 @@ {0x40b, 4, pc105, fi, NULL, NULL, Finnish}, {0x40c, 4, pc105, fr, NULL, NULL, French (Standard)}, {0x80c, 4, pc105, be, NULL, NULL, French (Belgian)}, +{0x410, 4, pc105, it, NULL, NULL, Italian}, {0x41d, 4, pc105, se, NULL, NULL, Swedish (Sweden)}, {-1, -1, NULL, NULL, NULL, NULL, NULL} }; --
Re: Transparent terminal window?
Turning on transparency in aterm is trivial and it worked well. I did experience the CPU problem as well, and the echo/refresh problem really made it unusable. -Simon - Original Message - From: Jean-Claude Gervais [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Saturday, April 05, 2003 4:00 AM Subject: RE: Transparent terminal window? Thanks Biju, Simon. Simon, you're right; ATerm builds easily on Cygwin, but it does have the display problem you mentioned. And on my system, with a very recent fresh install of Cygwin, it maxes out the CPU the second you start it. I haven't even tried to get it to display the transparent background though, have you? I will send a note to ATerm's author to tell him about the CPU bug. Thanks -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Simon Dong Sent: Friday, April 04, 2003 11:13 PM To: [EMAIL PROTECTED] Subject: Re: Transparent terminal window? Aterm doens't work properly on Cygwin/XFree86. It has a problem refreshing the screen. Whatever you type in will not be echoed until you move/resize the window. I got ETerm compile on Cygwin with transparent background. I couldn't get the background image to work (because I couldn't get imlib2 compiled with dll support, I think), but I really don't need it. I had to jump through a lot of hoops though, and didn't get the process documented. For what I can remember, I need to compile edb1.0.3 freetype-1.3.1 imlib2-1.0.6 And apply the kde patch to cygwin for shared mem support, which can be found at kde/cygwin's website. Then I can get Eterm built. -Simon - Original Message - From: Biju G C [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Friday, April 04, 2003 5:26 PM Subject: Re: Transparent terminal window? I have not seen any binary avialable on cygwin try to compile http://aterm.sourceforge.net/ for more screenshots http://xplanet.sourceforge.net/ cheers biju CyGNOME = Cygwin + GNOME http://cygnome.sourceforge.net/ --- Jean-Claude Gervais [EMAIL PROTECTED] wrote: Hi, I'm running WindowMaker, and I was wondering; does anyone know how to get transparent terminal windows? Can you do it with stock Cygwin-XFree or do you have to download and compile a specific terminal emulator? Thanks in advance. __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com
Re: Need some help on running the server on cygwin, and clients on FreeBSD.
This thread might be of interest to you... http://www.cygwin.com/ml/cygwin-xfree/2002-10/msg00091.html From: Constantine [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Need some help on running the server on cygwin, and clients on FreeBSD. Date: Thu, 03 Apr 2003 22:24:05 -0500 Hello! I am trying to run the X server on my cygwin machine, and the clients on my FreeBSD 4.7 box. As of today, to start the session I need to: 1. type startx 2. type xhost + 192.168.0.18 3. right-click, exit. 4. open ssh session, set display to 192.168.0.1, and run startkde. I do believe that this is a very complicated steps to follow, and I would highly appreciate if someone could explain, how could I set up the system to follow all these steps automatically. A bit of the reference material I was trying to follow: http://xfree86.cygwin.com/docs/ug/using-remote-session.html http://www.xfree86.org/current/Xserver.1.html#sect6 http://www.xfree86.org/current/xdm.1.html http://www.xs4all.nl/~zweije/xauth.html On both Cygwin (192.168.0.1) and FreeBSD (192.168.0.18) I use XFree86 4.2.0 (165). Thank you. Constantine. _ Help STOP SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail
Re: keyboard non-US
kaio wrote: winconfig.c: where? This was a messgae for Harold Hunt. You could apply this too, but this would require that you compile XWin yourself. If you just want the keyboard to be autodetected, just wait for the next test release. bye ago NP: VNV Nation - Structure -- [EMAIL PROTECTED] http://www.gotti.org ICQ: 126018723
Remote Graphical Acesss to SuSE from XP Continued
My great thanks to the quick reply to my message! I did have some success with the suggestion. However, I am still not able to view KDE running on my Linux server on my WinXP machine. Still, now have a new error message, but this is a good thing! It shows that Xfree86 is trying to get info from the server, which I was never able to confirm before. Same setup: WinXP 192.168.0.7 SuSE8.1 192.168.0.5 When I executed startxwin.sh, a new XFree86 window came up with the X cursor and a terminal window. From there I logged onto my linux machine (ssh -x 192.168.0.5) and echoed the display variable (echo $display). A blank line followed. Assuming that the variable was not set, I typed it in manually on the server and tried the process again. The following error was produced: Agent pid 1200 Xlib: connection to 192.168.0.5:0.0 refused by server Xlib: No protocol specified twm: unable to open display 192.168.0.5:0.0 Xlib: connection to 192.168.0.5:0.0 refused by server Xlib: No protocol specified xsetroot: unable to open display '192.168.0.5:0.0' My startxwin.bat file contains the following commands #! /bin/sh export DISPLAY=192.168.0.5:0.0 PATH=/usr/X11R6/bin:$PATH eval `ssh-agent` # Cleanup from last run. rm -rf /tmp/.X11-unix # Start the X Server. XWin -screen 0 1024 768 -noreset # Start the twm window manager. twm # Set a background color to hide that nasty X stipple. xsetroot -solid aquamarine4 # Start an xterm. xterm -n xterm-1 -j -ls -sb -sl 500 -rightbar -geometry +361+0 -e bash # Return from sh. exit Any information that anyone would be willing to provide would be very helpful! I thank you for your support of this message list. David White [EMAIL PROTECTED] - Reply: I don't see that you have started X? Meanwhile step by step.. getting an xterm.. | I start CygWin through the standard cygwin.bat file from installation. After starting cygwin.bat, type startxwin.sh in the shell-window.. | From the terminal window I can connect to 192.168.0.5 through TelNet | or SSH to change the DISPLAY variable. In my attempts I have set | DISPLAY to both 192.168.0.7:0.0 and 127.0.0.1:0.0. From the window opened in the X-window, do ssh -X 192.168.0.5 and there, do echo $DISPLAY (let ssh set the DISPLAY for you), if you get a value like 192.168.0.7:0.0 it should be working, here fire up a xterm. /Andy -- The eye of the linker rests on the libs! - Original Message: Sirs, I am attempting to access my SuSE 8.1 machine from WinXP by way of CygWin/XFree86 without success. I am able to access the Linux machine with SSH through a standard terminal window though. Unfortunately I am new to Linux, so I am going at this half blind. I have visited the message archives and have tried several suggested listed there, but I still have problems. My setup: (one network card each) Linux Machine: SuSE 8.1 running KDE 3.0 desktop all xfree86 packages installed IP 192.168.0.5 WinXP Machine: CygWin with all xfree86 and KDE2.2.2 packages installed IP 192.168.0.7 I start CygWin through the standard cygwin.bat file from installation. From the terminal window I can connect to 192.168.0.5 through TelNet or SSH to change the DISPLAY variable. In my attempts I have set DISPLAY to both 192.168.0.7:0.0 and 127.0.0.1:0.0. After I logout, I attempt set XHOST with XHOST +192.168.0.7. I get an error message; it is the only response I am ever able to get from XHOST with any IP address or host name entry. The error message is: xhost: unable to open display?? If I run xstart or xwin, a blank graphics window comes up with the X mouse cursor, etc, without an image. I have tried to setup different display variables in startxwin.bat, startxwin.sh, cygwin.bat and others without any change. The one interesting thing is that from a terminal window with in Xwin window I can type XHOST 192.168.0.5 and get an response saying that the address has been added. Still there is not any change in that Xwin window or any other one that is started from that terminal screen. I have followed all instructions I could find on the Internet about installation for CygWin and the KDE2.2.2 package (http://kde-cygwin.sourceforge.net/kde2/installation.php). Now I am at a loss. If someone could please offer ANY suggestions, it will be most appreciated!! Thank you for your time. David White whitedavida at hotmail dot com _ The new MSN 8: advanced junk mail protection and 2 months FREE* http://join.msn.com/?page=features/junkmail
Re: Remote Graphical Acesss to SuSE from XP Continued
| When I executed startxwin.sh, a new XFree86 window came up with the X | cursor and a terminal window. From there I logged onto my linux | machine (ssh -x | 192.168.0.5) and echoed the display variable (echo $display). A blank ssh -X, capital X to be used.. include a -v switch also to see verbosely what is happening. Uhh checking your startxwin.bat.. you do: (hmm is it bat, not shell-script?) #! /bin/sh export DISPLAY=192.168.0.5:0.0 PATH=/usr/X11R6/bin:$PATH [...] Have you edited it?? I mean 192.168.0.5.. that is the linux machine? And DISPLAY shouldn't be set to that machine, it should point to the local machine.. Start clean, with an unedited startxwn, to get started. /Andy p.s. for starters to quickly get around -- The eye of the linker rests on the libs!
src/winsup/cygwin ChangeLog
CVSROOT:/cvs/src Module name:src Branch: foo Changes by: [EMAIL PROTECTED] 2003-04-05 16:25:07 Modified files: winsup/cygwin : ChangeLog Log message: Patches: http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/winsup/cygwin/ChangeLog.diff?cvsroot=srconly_with_tag=foor1=1.1856r2=1.1856.2.1
Re: mkpasswd and mkgroup
--- Pierre A. Humblet [EMAIL PROTECTED] wrote: Corinna, Hmm. I feel like I'm eavesdropping. Following remarks made on the list this patch - allows to specify several domains at once with -d I just don't understand this one. Can someone be logged into multiple domains at once? - only prints SYSTEM and specials when the -l switch is given Isn't this a rather abrupt change from how these utilities have been working? Also, please do a patch for utils.sgml explaining the significance of the new functionality. __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com
Re: mkpasswd and mkgroup
At 08:27 AM 4/5/2003 -0800, Joshua Daniel Franklin wrote: --- Pierre A. Humblet [EMAIL PROTECTED] wrote: - allows to specify several domains at once with -d I just don't understand this one. Can someone be logged into multiple domains at once? You don't need to be logged in if the domains have a trust relationship (AFAIK). - only prints SYSTEM and specials when the -l switch is given Isn't this a rather abrupt change from how these utilities have been working? Nobody should ever populate /etc/passwd with -d without having a -l (either concurrent or in a separate call), so SYSTEM will be there. The current program causes duplicate entries because SYSTEM (and others) are added in each call. People have reported writing scripts to remove duplicates, so I am just trying to be nice and avoid having them. Also, please do a patch for utils.sgml explaining the significance of the new functionality. Sure. Pierre
RE: 1.3.22: TMPDIR Initialization
On Sat, 2003-04-05 at 17:43, Mark Moore wrote: Christopher, I've read the http://cygwin.com/lists.html. Here's the description for the mailing list I posted to: ... The passage [cygwin-apps] is the preferred location for design discussions and bug reports regarding cygwin's setup.exe program prompted me to send the post to the cygwin-apps forum. If this is no longer the case, someone may want to change the instructions. What you posted has nothing to do with setup.exe. Unless you're familiar with setup.exe, it's an honest mistake - he saw what he thought was a setup problem, and in the Windows world, this generally means a problem with the installer itself. -Samrobb -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
nano and TERM
Possibly this is a slight simplification, but broadly speaking is the following an accurate description of things? Assuming all necessary paths exist, Windows executables (fc, for instance; can't find much else that's any use) can be called from within Cygwin; and vice-versa (ls, md5sum, diff; any number of useful others). That's why find, sort, etc can be problematic, if a user ends up calling the alternative version to that required. Lately (can't remember when it started) I've been unable to use nano from outside Cygwin: I get a Error opening terminal: cygwin message. (On the other hand, pico works fine.) Please can you tell me whether this is a minor correct-able glitch in nano or whether there's a deeper problem (like, I've misunderstood the dual-use principle described above). Thank you. Fergus -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl 5.8.0 debugger requires double ENTER key to act on commands: cygwin 1.3.22
Hallo Gavin, you wrote: After upgrading from cygwin 1.3.20 to cygwin 1.3.22, the new version of Perl (5.8.0) appears to require 2 ENTER keys to be hit for any command to be acted on by the Perl debugger. The previous version of Perl (5.7.x) on cygwin 1.3.20 did not have this problem. This was already reported, see here: http://cygwin.com/ml/cygwin/2003-03/msg02253.html http://cygwin.com/ml/cygwin/2003-03/msg02272.html Any suggestions on how to resolve this problem? See the second link above, try to set: $ export PERLIO=perlio in your environment. Besides that you may file a bugreport using the perlbug script. Gerrit -- =^..^= -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl-5.8.0-2 CYGWIN1.DLL:setlinebuf failure
Hallo Greg, And now running perldoc, I only get 2 error messages instead of 4 about invalid page faults in cygwin DLLs. Perhaps because the perldoc page headers seem to have gone. What error messages are you getting when you run perldoc? I doon't get error messages when running it. But installing modules, the 4-5 batches of 10 or 11 error messages during the make and make install stages and the batches of 2 error messages during the make test stage remain as they were in 5.8.0-1. I don't see error messages here, can you give some more details about this, please? Gerrit -- =^..^= -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: PERL 5.8.0-2 and Haase_Gerrit/1.3/perl/Tk800.023-1.patch - taking this to Tk800.024
Hallo Alan, The PERL package Tk800.023 is now Tk800.024 but Haase Gerrit's patch to get this package to work under Cygwin only works to Tk800.023. Since I wanted to use Tk800.024, I had to figure out how to do this. It was reported that Tk800.024 builds without patching it or with minimal tweaks like the one you need to apply to get it compiled after updating 023 w/ patch to 024. Gerrit -- =^..^= -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: tin says Can't get a (fully-qualified) domain-name!
Yadin Y. Goldschmidt wrote: create a file /etc/tin/tin.defaults and put a line there domainname=yourdomain If you can read this then you can assume that that fixed the problem, thanks! I also had to set disable_sender=ON in the same file (see http://www.math.psu.edu/local_doc/tin/tin.defaults for a description of tin.defaults though I suspect it can be found elsewhere as well). -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl 5.8.0-2 and WaitForMultipleObjects under win98
Hallo Nicholas, 307776554 [main] perl 614369 sync_with_child: WaitForMultipleObjects timed out I believe this to be a rebasing issue. I honestly think this is the result of fiddling with the stack size during the build (--stack,8388608) and non NT systems just don't like it. I may be wrong... Alright, I'll build a perl without increasing the stack for you, and you can test it, so we'll see if it makes a difference. I'll send you a private mail with the URL after finishing the build. Gerrit -- =^..^= -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl 5.8.0-2 and WaitForMultipleObjects under win98
Hallo Nicholas, Pierre, This is also happening on WinME when I try to run either of those programs with the new perl. Except I get something Do you or someone else with Win98 or WinME have some time to do a complete build of perl, including the full testsuite? I.e. unpack the source tarball and cd perl-5.8.0-2 and run the build.sh script and send me all the logfiles created during the build tests, please! Gerrit -- =^..^= -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: 1.3.22: TMPDIR Initialization
Robert, If the permission on the folder pointed to by $TEMP is going to be changed during install, it seems to me it would be done by setup.exe. If not, where would that operation be performed? It looks like chmod doesn't cut it: [EMAIL PROTECTED] ~ $ ls -ld $TEMP d---rwx---+ 7 Administ SYSTEM 16384 Apr 5 00:10 /cygdrive/c/WINNT/TEMP [EMAIL PROTECTED] ~ $ chmod +rw $TEMP [EMAIL PROTECTED] ~ $ ls -ld $TEMP drw-rwxr--+ 7 Administ SYSTEM 16384 Apr 5 00:10 /cygdrive/c/WINNT/TEMP [EMAIL PROTECTED] ~ $ perldoc perldoc Error in tempfile() using /XX: Parent directory (/) is not writable at /usr/bin/perldoc line 564 [EMAIL PROTECTED] ~ $ Notice that global rw permissions don't stick. -Original Message- From: Robert Collins [mailto:[EMAIL PROTECTED] Sent: Friday, April 04, 2003 11:51 PM To: Mark Moore Cc: [EMAIL PROTECTED] Subject: RE: 1.3.22: TMPDIR Initialization On Sat, 2003-04-05 at 17:43, Mark Moore wrote: Christopher, I've read the http://cygwin.com/lists.html. Here's the description for the mailing list I posted to: ... The passage [cygwin-apps] is the preferred location for design discussions and bug reports regarding cygwin's setup.exe program prompted me to send the post to the cygwin-apps forum. If this is no longer the case, someone may want to change the instructions. What you posted has nothing to do with setup.exe. Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: nano and TERM
Hmm, unable to reproduce your problem here so far, prehaps some more details would be useful. Regards, Gareth - nano packaging type guy. - Original Message - From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Sent: Saturday, April 05, 2003 6:05 PM Subject: nano and TERM Possibly this is a slight simplification, but broadly speaking is the following an accurate description of things? Assuming all necessary paths exist, Windows executables (fc, for instance; can't find much else that's any use) can be called from within Cygwin; and vice-versa (ls, md5sum, diff; any number of useful others). That's why find, sort, etc can be problematic, if a user ends up calling the alternative version to that required. Lately (can't remember when it started) I've been unable to use nano from outside Cygwin: I get a Error opening terminal: cygwin message. (On the other hand, pico works fine.) Please can you tell me whether this is a minor correct-able glitch in nano or whether there's a deeper problem (like, I've misunderstood the dual-use principle described above). Thank you. Fergus -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/ -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: 1.3.22: TMPDIR Initialization
[Quoting adjusted] * Mark Moore (03-04-05 10:58 +0100) The passage [cygwin-apps] is the preferred location for design discussions and bug reports regarding cygwin's setup.exe program prompted me to send the post to the cygwin-apps forum. What you posted has nothing to do with setup.exe. If the permission on the folder pointed to by $TEMP is going to be changed during install, it seems to me it would be done by setup.exe. If not, where would that operation be performed? The postinstall script delivered with the package. Thorsten -- Content-Type: text/explicit; charset=ISO-8859-666 (Parental Advisory) Content-Transfer-Warning: message contains innuendos not suited for children under the age of 18 -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
ORB
Hello, I need a ORB for an application under Cygwin. I know Mico and Omniorb but are there other ? Etienne I -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
perldoc prints control characters
If I run the following command: $ perldoc Getopt::Long Then I get too printed the nasty control characters, e.g.: ESC[1mNAMEESC[0m Getopt::Long - Extended processing of command line options I am not sure if this is related to perl 5.8.0-2 I looked into my environment but there is nothing special. Also the programs 'man' and 'less' works IMHO fine. Harald -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl-5.8.0-2 CYGWIN1.DLL:setlinebuf failure
On Sat, 05 Apr 2003, Gerrit P. Haase wrote: Hallo Greg, And now running perldoc, I only get 2 error messages instead of 4 about invalid page faults in cygwin DLLs. Perhaps because the perldoc page headers seem to have gone. What error messages are you getting when you run perldoc? I doon't get error messages when running it. The same messages I had been getting for a couple of months with 5.8.0-1. As I posted here: Linkname: Greg Matheson - perl error messages with cygwin 1.3.18-1 URL: http://cygwin.com/ml/cygwin/2003-01/msg00028.html This is on Windows 98. And now with cygwin 1.3.22-1. -- Greg MathesonThere is nothing as good as a Dr Bean's Penpal Poolpractical theory. Address: palpool --Kurt Lewin's practicality maxim corollary Domain: @cn91.chinmin.edu.tw -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: 1.3.22: TMPDIR Initialization
On Sat, 2003-04-05 at 18:58, Mark Moore wrote: Robert, If the permission on the folder pointed to by $TEMP is going to be changed during install, it seems to me it would be done by setup.exe. If not, where would that operation be performed? Setup doesn't set any special permissions. It sets a generic umask and mode, and applies that throughout. ALL special permissions, scripts, and local customisations are done via post-install scripts, which run after setup extracts the package files, and before setup completes. So, if the TEMP permissions need adjusting, it's the job of a postinstall script (perhaps base-files in this case) to do that. Cheers, Rob -- GPG key available at: http://users.bigpond.net.au/robertc/keys.txt. signature.asc Description: This is a digitally signed message part
Re: ORB
Hello, * On Fri, Apr 04, 2003 at 05:07:50PM +0200, Vaillant Etienne [EMAIL PROTECTED] wrote: I need a ORB for an application under Cygwin. I know Mico and Omniorb but are there other ? TAO may be -- based on the C++ library ACE. I am not sure if it has been ported for Cygwin, but it is available on many different systems -- unices, Windows, etc. HTH, -- Luc Hermitte -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl 5.8.0-2 and WaitForMultipleObjects under win98
[EMAIL PROTECTED] wrote: Hallo Nicholas, Pierre, This is also happening on WinME when I try to run either of those programs with the new perl. Except I get something Do you or someone else with Win98 or WinME have some time to do a complete build of perl, including the full testsuite? I.e. unpack the source tarball and cd perl-5.8.0-2 and run the build.sh script and send me all the logfiles created during the build tests, please! Sure, I'll try. Hopefully it won't require rebase to run the tests... Cheers, Nicholas -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Help -- Cygwin configuration problem
I'm redirecting you to [EMAIL PROTECTED] Please research their archives. Please remove [EMAIL PROTECTED] from your responses. Earnie. Wei Qin wrote: I have some trouble running the generated configure script on Cygwin, it complains that /bin/sh: not found at various occasions. I created a link from /bin/sh.exe to /bin.sh, but useless. Please help out if you know how to fix this. Thanks in advance. Wei -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl-5.8.0-2 CYGWIN1.DLL:setlinebuf failure
Hallo Greg, What error messages are you getting when you run perldoc? I doon't get error messages when running it. The same messages I had been getting for a couple of months with 5.8.0-1. As I posted here: Linkname: Greg Matheson - perl error messages with cygwin 1.3.18-1 URL: http://cygwin.com/ml/cygwin/2003-01/msg00028.html This is on Windows 98. And now with cygwin 1.3.22-1. Thanks, I remember now. I have no idea what the problem is though. Gerrit -- =^..^= -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: PERL 5.8.0-2 and Haase_Gerrit/1.3/perl/Tk800.023-1.patch - taking this to Tk800.024
Gerrit, Thanks for the email. I have been using CYGWIN for about 3/4 years now, both at home and at work, and I needed to get this item to work using cygwin perl. Now it does. It was reported that Tk800.024 builds without patching it or with minimal tweaks like the one you need to apply to get it compiled after updating 023 w/ patch to 024. The libraries compile (except for the warning below), but one of the test programs failed to link due to the undefined reference to `_win32_get_osfhandle' issue in Tk800.024-1/Event/Event.xs. I found the answer on how to fix that in http://sources.redhat.com/ml/cygwin/2002-09/msg00752.html. To build I execute the command perl Makefile.PL make make test make install. Hence my reason for patching the source. My purpose in sending the email was to ensure that nobody else had to re-invent the wheel. Maybe you could try this patch out, and then release it for the cygwin community. Since emailing the group, I have fixed the below problem: /usr/include/limits.h:143:1: warning: this is the location of the previous definition gcc -c -I.. -I../pTk/mTk/xlib -I. -Ibitmaps -I../pTk/mTk/xlib -DPERL_USE_SA FE_PUTENV -fno-strict-aliasing -DUSEIMPORTLIB -O2 -DVERSION=\800.024\ -D XS_VERSION=\800.024\ -I/usr/lib/perl5/5.8.0/cygwin-multi-64int/CORE -D__WIN32__ -D_WIN32 -DWIN 32 -Wall -Wno-implicit-int -Wno-comment -Wno-unused -D__USE_FIXED_PROTOTYPES __ imgGIF.c In file included from tkPort.h:28, from tk.h:71, from imgGIF.c:35: tkWinPort.h:54:1: warning: OPEN_MAX redefined In file included from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/include/limits.h:132, from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/include/syslimits.h:7, from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/include/limits.h:11, from tkWinPort.h:31, from tkPort.h:28, from tk.h:71, from imgGIF.c:35: Hence my current (complete) patch is (which is largely based on your original patch): Start_Of_Patch diff -r -u Tk800.024/Event/Event.xs Tk800.024-1/Event/Event.xs --- Tk800.024/Event/Event.xs2002-01-07 03:35:44.0 -0600 +++ Tk800.024-1/Event/Event.xs 2003-04-04 13:49:42.0 -0600 @@ -4,6 +4,17 @@ modify it under the same terms as Perl itself. */ +/* + * Fix required to fix: undefined reference to `_win32_get_osfhandle' issue as directed in: + * + * http://sources.redhat.com/ml/cygwin/2002-09/msg00752.html + * + */ +#ifdef __CYGWIN__ +# define win32_get_osfhandle _get_osfhandle +# define win32_open_osfhandle _open_osfhandle +#endif + #include EXTERN.h #include perl.h #include XSUB.h diff -r -u Tk800.024/Makefile.PL Tk800.024-1/Makefile.PL --- Tk800.024/Makefile.PL 2002-03-05 08:00:43.0 -0600 +++ Tk800.024-1/Makefile.PL 2003-04-04 13:49:42.0 -0600 @@ -154,7 +154,7 @@ } sub MY::postamble { -' + my $str = ' html : subdirs manifypods @cd pod $(MAKE) html $(PASTHRU) @@ -184,6 +184,8 @@ $(MAKE) manifest '; + $str =~ s/DEFINE=.*// if($^O eq 'cygwin'); + $str; } diff -r -u Tk800.024/README.cygwin Tk800.024-1/README.cygwin --- Tk800.024/README.cygwin 2000-09-08 04:27:27.0 -0500 +++ Tk800.024-1/README.cygwin 2003-04-04 13:49:42.0 -0600 @@ -90,7 +90,7 @@ (sometimes $^O eq 'MSWin32' is used instead of $Tk::platform eq 'MSWin32' assuming they are equivalent and vice versa) + #define distinctions -_WIN32defined by gcc, if X11 need to undef +_WIN32win_arch MSWin32 __WIN32__ win_arch MSWin32, also defined in pTk/Lang.h WIN32 from #include windows.h when precedes #include perl.h then need @@ -107,7 +107,7 @@ README.cygwin MANIFEST * documentation -Makefile.PL Tk/MMutil.pm pTk/Makefile.PL +Makefile.PL Tk/MMutil.pm pTk/Makefile.PL myConfig * make stuff pTk/mTk/win/tkWinX.c @@ -149,6 +149,6 @@ + Fix gcc warnings, mostly in pTk win32 flavor? + Test with cygwin-xfree X Server? -Thu Aug 31 12:54:09 BST 2000 +Sun Jun 10 21:34:30 PST 2001 Eric Fifer [EMAIL PROTECTED] [EMAIL PROTECTED] diff -r -u Tk800.024/Tk/MMutil.pm Tk800.024-1/Tk/MMutil.pm --- Tk800.024/Tk/MMutil.pm 2002-03-05 08:27:40.0 -0600 +++ Tk800.024-1/Tk/MMutil.pm2003-04-04 13:49:42.0 -0600 @@ -238,13 +238,10 @@ push(@def,qw(-D__CYGWIN__)); if ($win_arch eq 'MSWin32') { - push(@def,qw(-D__WIN32__)) unless $self-{'DEFINE'} =~ /-D__WIN32__/; + push(@def,qw(-D__WIN32__ -D_WIN32)) + unless $self-{'DEFINE'} =~ /-D__WIN32__/; push(@def,qw(-DWIN32)) if $self-{'NAME'} eq 'Tk::pTk'; } - elsif ($win_arch eq 'x') - { - push(@def,qw(-U_WIN32)); - } } foreach (@inc) { @@ -519,16 +516,12 @@ push(@opt,'LD' = 'gcc -shared'); if ($win_arch eq
Re: perldoc prints control characters
Gerrit P. Haase [EMAIL PROTECTED] writes: In perl-5.8.0-2 perldoc is patched to always use the -t switch, your perldoc is not the perldoc from perl-5.8.0-2. Uups, you are right. It was a 4 year old perldoc from /usr/local/bin. I have cleaned up the files and now all works fine. Thanks for the hint. Harald -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: PERL 5.8.0-2 and Haase_Gerrit/1.3/perl/Tk800.023-1.patch - taking this to Tk800.024
Sorry - my original reply went to Gerrit only - having read on the list that it should be going to the group, I am sending this to the group ... Gerrit, Thanks for the email. I have been using CYGWIN for about 3/4 years now, both at home and at work, and I needed to get this item to work using cygwin perl. Now it does. It was reported that Tk800.024 builds without patching it or with minimal tweaks like the one you need to apply to get it compiled after updating 023 w/ patch to 024. The libraries compile (except for the warning below), but one of the test programs failed to link due to the undefined reference to `_win32_get_osfhandle' issue in Tk800.024-1/Event/Event.xs. I found the answer on how to fix that in http://sources.redhat.com/ml/cygwin/2002-09/msg00752.html. To build I execute the command perl Makefile.PL make make test make install. Hence my reason for patching the source. My purpose in sending the email was to ensure that nobody else had to re-invent the wheel. Maybe you could try this patch out, and then release it for the cygwin community. Since emailing the group, I have fixed the below problem: /usr/include/limits.h:143:1: warning: this is the location of the previous definition gcc -c -I.. -I../pTk/mTk/xlib -I. -Ibitmaps -I../pTk/mTk/xlib -DPERL_USE_SA FE_PUTENV -fno-strict-aliasing -DUSEIMPORTLIB -O2 -DVERSION=\800.024\ -D XS_VERSION=\800.024\ -I/usr/lib/perl5/5.8.0/cygwin-multi-64int/CORE -D__WIN32__ -D_WIN32 -DWIN 32 -Wall -Wno-implicit-int -Wno-comment -Wno-unused -D__USE_FIXED_PROTOTYPES __ imgGIF.c In file included from tkPort.h:28, from tk.h:71, from imgGIF.c:35: tkWinPort.h:54:1: warning: OPEN_MAX redefined In file included from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/include/limits.h:132, from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/include/syslimits.h:7, from /usr/lib/gcc-lib/i686-pc-cygwin/3.2/include/limits.h:11, from tkWinPort.h:31, from tkPort.h:28, from tk.h:71, from imgGIF.c:35: Hence my current (complete) patch is (which is largely based on your original patch): Start_Of_Patch diff -r -u Tk800.024/Event/Event.xs Tk800.024-1/Event/Event.xs --- Tk800.024/Event/Event.xs2002-01-07 03:35:44.0 -0600 +++ Tk800.024-1/Event/Event.xs 2003-04-04 13:49:42.0 -0600 @@ -4,6 +4,17 @@ modify it under the same terms as Perl itself. */ +/* + * Fix required to fix: undefined reference to `_win32_get_osfhandle' issue as directed in: + * + * http://sources.redhat.com/ml/cygwin/2002-09/msg00752.html + * + */ +#ifdef __CYGWIN__ +# define win32_get_osfhandle _get_osfhandle +# define win32_open_osfhandle _open_osfhandle +#endif + #include EXTERN.h #include perl.h #include XSUB.h diff -r -u Tk800.024/Makefile.PL Tk800.024-1/Makefile.PL --- Tk800.024/Makefile.PL 2002-03-05 08:00:43.0 -0600 +++ Tk800.024-1/Makefile.PL 2003-04-04 13:49:42.0 -0600 @@ -154,7 +154,7 @@ } sub MY::postamble { -' + my $str = ' html : subdirs manifypods @cd pod $(MAKE) html $(PASTHRU) @@ -184,6 +184,8 @@ $(MAKE) manifest '; + $str =~ s/DEFINE=.*// if($^O eq 'cygwin'); + $str; } diff -r -u Tk800.024/README.cygwin Tk800.024-1/README.cygwin --- Tk800.024/README.cygwin 2000-09-08 04:27:27.0 -0500 +++ Tk800.024-1/README.cygwin 2003-04-04 13:49:42.0 -0600 @@ -90,7 +90,7 @@ (sometimes $^O eq 'MSWin32' is used instead of $Tk::platform eq 'MSWin32' assuming they are equivalent and vice versa) + #define distinctions -_WIN32defined by gcc, if X11 need to undef +_WIN32win_arch MSWin32 __WIN32__ win_arch MSWin32, also defined in pTk/Lang.h WIN32 from #include windows.h when precedes #include perl.h then need @@ -107,7 +107,7 @@ README.cygwin MANIFEST * documentation -Makefile.PL Tk/MMutil.pm pTk/Makefile.PL +Makefile.PL Tk/MMutil.pm pTk/Makefile.PL myConfig * make stuff pTk/mTk/win/tkWinX.c @@ -149,6 +149,6 @@ + Fix gcc warnings, mostly in pTk win32 flavor? + Test with cygwin-xfree X Server? -Thu Aug 31 12:54:09 BST 2000 +Sun Jun 10 21:34:30 PST 2001 Eric Fifer [EMAIL PROTECTED] [EMAIL PROTECTED] diff -r -u Tk800.024/Tk/MMutil.pm Tk800.024-1/Tk/MMutil.pm --- Tk800.024/Tk/MMutil.pm 2002-03-05 08:27:40.0 -0600 +++ Tk800.024-1/Tk/MMutil.pm2003-04-04 13:49:42.0 -0600 @@ -238,13 +238,10 @@ push(@def,qw(-D__CYGWIN__)); if ($win_arch eq 'MSWin32') { - push(@def,qw(-D__WIN32__)) unless $self-{'DEFINE'} =~ /-D__WIN32__/; + push(@def,qw(-D__WIN32__ -D_WIN32)) + unless $self-{'DEFINE'} =~ /-D__WIN32__/; push(@def,qw(-DWIN32)) if $self-{'NAME'} eq 'Tk::pTk'; } - elsif ($win_arch eq 'x') - { -
RE: WinXP username with spaces -- wmaker ko
All, I have the same problem (on my machine at home) - However I do have a (partial) solution (I am still working on the final solution): This has to be done after CYGWIN installation, but BEFORE and user logon. Edit the /etc/passwd file and replace the space in the first field, and edit the field /home/something with space to /home/something_with_out_space and save. I am working on a scripting solution to install as part of my installation - maybe this could be incorporated as part of the main CYGWIN release. Below is a script fragment on how I am proposing to automate this fix. ThePasswordFile=/etc/passwd ModUserName=$(/usr/bin/echo ${USER} | /usr/bin/sed -e 's/ /_/g') ... ModHomeUserName=$(/usr/bin/basename ${HOME} | /usr/bin/sed -e 's/ /_/g') ModHomeUserDirName=$(/usr/bin/dirname ${HOME})/ ModHome=${ModHomeUserDirName}${ModHomeUserName} /usr/bin/rm -f ${ThePasswordFile} if [ ! -e ${ThePasswordFile} -a ! -L ${ThePasswordFile} ] then /usr/bin/mkpasswd -l -c | /usr/bin/gawk -F':' -v oldUser=${USER} -v newUser=${ModUserName} -v newUserHomeDir=${ModHome} ' BEGIN {OFS=: ; } $1 ~ oldUser {$1=newUser ; $6=newUserHomeDir; } {print $0 ; } ' ${ThePasswordFile} fi Hope this helps. Alan -Original Message- From: kaio [mailto:[EMAIL PROTECTED] Sent: April 4, 2003 10:48 To: [EMAIL PROTECTED] Subject: WinXP username with spaces -- wmaker ko i have my administrator user named with a space on WInXP. cygwin installation is ok, but windowmaker don't install: seems that the space generates a new parameter in installation commands! in fact, wmaker installation creates a new dir in /home/ with the first half of username! -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Bug in GUI App WinMain Arguements
On Fri, Apr 04, 2003 at 07:58:27PM +0100, Max Bowsher wrote: BB wrote: Also, according to the Cygwin Users Guide, I have included the following line in the code. WinMainCRTStartup() { mainCRTStartup(); } It doesnt seem to be called. Is the User Guide correct? The same problem occurs if I omit the line and use the -e _mainCRTStartup as suggested in the User Guide. This sounds like out-of-date info to me (but that's just a guess). OK, I'd like to make sure this is up-to-date, but I've never built a GUI Win32 application in Cygwin (well, except setup.exe). Can someone post a helloworld-gui example or something? Or, here's what the User's Guide currently says, please reply with inaccuracies: GUI Mode Applications Cygwin allows you to build programs with full access to the standard Windows 32-bit API, including the GUI functions as defined in any Microsoft or off-the-shelf publication. However, the process of building those applications is slightly different, as you'll be using the GNU tools instead of the Microsoft tools. For the most part, your sources won't need to change at all. However, you should remove all __export attributes from functions and replace them like this: int foo (int) __attribute__ ((__dllexport__)); int foo (int i) For most cases, you can just remove the __export and leave it at that. For convenience sake, you might want to include the following code snippet when compiling GUI programs. If you don't, you will want to add -e _mainCRTStartup to your link line in your Makefile. #ifdef __CYGWIN__ WinMainCRTStartup() { mainCRTStartup(); } #endif The Makefile is similar to any other UNIX-like Makefile, and like any other Cygwin makefile. The only difference is that you use gcc -mwindows to link your program into a GUI application instead of a command-line application. Here's an example: myapp.exe : myapp.o myapp.res gcc -mwindows myapp.o myapp.res -o $@ myapp.res : myapp.rc resource.h windres $ -O coff -o $@ Note the use of windres to compile the Windows resources into a COFF-format .res file. That will include all the bitmaps, icons, and other resources you need, into one handy object file. Normally, if you omitted the -O coff it would create a Windows .res format file, but we can only link COFF objects. So, we tell windres to produce a COFF object, but for compatibility with the many examples that assume your linker can handle Windows resource files directly, we maintain the .res naming convention. For more information on windres, consult the Binutils manual. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: Bash and xterm window title
Ajay, The preferred answer is: Read the BASH manual, where the details of how to use the special escape codes recognized in the PS1 string are fully explained. To wit: -==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==- PROMPTING When executing interactively, bash displays the primary prompt PS1 when it is ready to read a command, and the secondary prompt PS2 when it needs more input to complete a command. Bash allows these prompt strings to be customized by inserting a number of backslash-escaped special characters that are decoded as follows: \a an ASCII bell character (07) \d the date in Weekday Month Date format (e.g., Tue May 26) \D{format} the format is passed to strftime(3) and the result is inserted into the prompt string; an empty format results in a locale-specific time representation. The braces are required \e an ASCII escape character (033) \h the hostname up to the first `.' \H the hostname \j the number of jobs currently managed by the shell \l the basename of the shell's terminal device name \n newline \r carriage return \s the name of the shell, the basename of $0 (the portion following the final slash) \t the current time in 24-hour HH:MM:SS format \T the current time in 12-hour HH:MM:SS format \@ the current time in 12-hour am/pm format \A the current time in 24-hour HH:MM format \u the username of the current user \v the version of bash (e.g., 2.00) \V the release of bash, version + patchelvel (e.g., 2.00.0) \w the current working directory \W the basename of the current working directory \! the history number of this command \# the command number of this command \$ if the effective UID is 0, a #, otherwise a $ \nnn the character corresponding to the octal number nnn \\ a backslash \[ begin a sequence of non-printing characters, which could be used to embed a terminal control sequence into the prompt \] end a sequence of non-printing characters -==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==- I believe you're interested in the \w and / or \W sequences. Randall Schulz At 08:08 2003-04-05, you wrote: Hi, I had some problem with a recent version of tcsh and so I'm using bash. The only problem I have with bash is that it re-writes the window title with the current working directory. I use a script to name(title) my xterms and bash overwrites it. tcsh does not do that. I set PS1=$ and now it doesn't do it. The original PS1 is set to: $ echo $PS1 \[\033]0;\w\007 [EMAIL PROTECTED] \[\033[33m\w\033[0m\] $ The desired behavior in my case is just the prompt gets updated with the PWD and not the window title. How can I do this? TIA, -ajay -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: PERL 5.8.0-2 and Haase_Gerrit/1.3/perl/Tk800.023-1.patch - taking this to Tk800.024
On Sat, 5 Apr 2003, Alan Miles wrote: Sorry - my original reply went to Gerrit only - having read on the list that it should be going to the group, I am sending this to the group ... [snip] Alan, Did you by any chance notice which return address Gerrit used? ;-) FYI, your original reply is here: http://cygwin.com/ml/cygwin/2003-04/msg00502.html Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: PERL 5.8.0-2 and Haase_Gerrit/1.3/perl/Tk800.023-1.patch - taking this to Tk800.024
Igor, I am using MS Outlook 2000 - Gerrit's email was: [EMAIL PROTECTED] when I checked the properties of this email I got: [EMAIL PROTECTED] Thus I am guessing I inadvertently sent duplicate messages to the group - my apologies for that. Alan -Original Message- From: Igor Pechtchanski [mailto:[EMAIL PROTECTED] Sent: April 5, 2003 10:41 To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: RE: PERL 5.8.0-2 and Haase_Gerrit/1.3/perl/Tk800.023-1.patch - taking this to Tk800.024 On Sat, 5 Apr 2003, Alan Miles wrote: Sorry - my original reply went to Gerrit only - having read on the list that it should be going to the group, I am sending this to the group ... [snip] Alan, Did you by any chance notice which return address Gerrit used? ;-) FYI, your original reply is here: http://cygwin.com/ml/cygwin/2003-04/msg00502.html Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: WinXP username with spaces -- wmaker ko
Alan, The fix below has a few problems (one of which is unconditionally removing /etc/passwd). IMO, it's better to fix the scripts that have improper quoting, or, barring that, unset HOME in your Windows environment (or modify /etc/profile) and make sure your username doesn't contain a space. Going out of your way to accomodate scripts without proper quoting will only lead to the proliferation of such scripts. Igor On Sat, 5 Apr 2003, Alan Miles wrote: All, I have the same problem (on my machine at home) - However I do have a (partial) solution (I am still working on the final solution): This has to be done after CYGWIN installation, but BEFORE and user logon. Edit the /etc/passwd file and replace the space in the first field, and edit the field /home/something with space to /home/something_with_out_space and save. I am working on a scripting solution to install as part of my installation - maybe this could be incorporated as part of the main CYGWIN release. Below is a script fragment on how I am proposing to automate this fix. ThePasswordFile=/etc/passwd ModUserName=$(/usr/bin/echo ${USER} | /usr/bin/sed -e 's/ /_/g') ... ModHomeUserName=$(/usr/bin/basename ${HOME} | /usr/bin/sed -e 's/ /_/g') ModHomeUserDirName=$(/usr/bin/dirname ${HOME})/ ModHome=${ModHomeUserDirName}${ModHomeUserName} /usr/bin/rm -f ${ThePasswordFile} if [ ! -e ${ThePasswordFile} -a ! -L ${ThePasswordFile} ] then /usr/bin/mkpasswd -l -c | /usr/bin/gawk -F':' -v oldUser=${USER} -v newUser=${ModUserName} -v newUserHomeDir=${ModHome} ' BEGIN {OFS=: ; } $1 ~ oldUser {$1=newUser ; $6=newUserHomeDir; } {print $0 ; } ' ${ThePasswordFile} fi Hope this helps. Alan -Original Message- From: kaio [mailto:[EMAIL PROTECTED] Sent: April 4, 2003 10:48 To: [EMAIL PROTECTED] Subject: WinXP username with spaces -- wmaker ko i have my administrator user named with a space on WInXP. cygwin installation is ok, but windowmaker don't install: seems that the space generates a new parameter in installation commands! in fact, wmaker installation creates a new dir in /home/ with the first half of username! -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: WinXP username with spaces -- wmaker ko
Igor, Granted - There are issues with the solution - that is why I said partial solution ( I need to resolve ALL issues) The challenge really has to do with /usr/bin/mkpasswd (which gets called by /etc/postinstall/passwd-grp.sh during the install process) From what I can tell, it is the item that generates the password file, which affects HOME and the home directory setup. Often, users don't have access to change their id's, so there Windows logon name may have spaces in them. We cannot chnage that. Maybe I wasn't too clear in my thoughts - until I have a scripted solution (the code fragment I presented), I manually fix the problem by: 1/ Install CYGWIN (say to c:\cygwin) - I install the icon to my desktop and start menu 2/ Using Windows Wordpad, open the /etc/passwd file (e.g., c:\cygwin\etc\passwd) 3/ Change field 1 (the user name) - change any spaces to _ characters (i.e., change Alan Miles to Alan_Miles) 4/ Change field 6 (the home directory setting) - change any spaces to _ characters (i.e., change /home/Alan Miles to /home/Alan_Miles) 5/ Save /etc/passwd file (e.g., c:\cygwin\etc\passwd) 6/ Click on the cygwin desktop icon to logon to cygwin for the first time. /etc/profile creates the home directory the first time: (fragment from /etc/profile: # If the home directory doesn't exist, create it. if [ ! -d $HOME ]; then mkdir -p $HOME # copy skeleton files cp -a /etc/skel/. $HOME fi ) Now the cygwin creates the home directory properly (i.e., /home/Alan_Miles), HOME is set to /home/Alan_Miles, and USER is set to Alan_Miles I have not had any trouble with this manual setup. This is the item I would like to either fix by scripting, e.g., the partial code fragment already presented, or better still patch /usr/bin/mkpasswd to convert any names it sees with spaces in them to _ characters. Then problem completely solved. I appreciate your input. BTW, Igor got your other message regarding Perl - you said: This wasn't meant to chide you; You weren't chiding me - my first message I thought went to Gerrit only and should have gone to the list (which is why I then resent it). All you did was point out that I inadvertently duplicated messages - no problem. Alan -Original Message- From: Igor Pechtchanski [mailto:[EMAIL PROTECTED] Sent: April 5, 2003 10:53 To: [EMAIL PROTECTED] Cc: kaio; [EMAIL PROTECTED] Subject: RE: WinXP username with spaces -- wmaker ko Alan, The fix below has a few problems (one of which is unconditionally removing /etc/passwd). IMO, it's better to fix the scripts that have improper quoting, or, barring that, unset HOME in your Windows environment (or modify /etc/profile) and make sure your username doesn't contain a space. Going out of your way to accomodate scripts without proper quoting will only lead to the proliferation of such scripts. Igor On Sat, 5 Apr 2003, Alan Miles wrote: All, I have the same problem (on my machine at home) - However I do have a (partial) solution (I am still working on the final solution): This has to be done after CYGWIN installation, but BEFORE and user logon. Edit the /etc/passwd file and replace the space in the first field, and edit the field /home/something with space to /home/something_with_out_space and save. I am working on a scripting solution to install as part of my installation - maybe this could be incorporated as part of the main CYGWIN release. Below is a script fragment on how I am proposing to automate this fix. ThePasswordFile=/etc/passwd ModUserName=$(/usr/bin/echo ${USER} | /usr/bin/sed -e 's/ /_/g') ... ModHomeUserName=$(/usr/bin/basename ${HOME} | /usr/bin/sed -e 's/ /_/g') ModHomeUserDirName=$(/usr/bin/dirname ${HOME})/ ModHome=${ModHomeUserDirName}${ModHomeUserName} /usr/bin/rm -f ${ThePasswordFile} if [ ! -e ${ThePasswordFile} -a ! -L ${ThePasswordFile} ] then /usr/bin/mkpasswd -l -c | /usr/bin/gawk -F':' -v oldUser=${USER} -v newUser=${ModUserName} -v newUserHomeDir=${ModHome} ' BEGIN {OFS=: ; } $1 ~ oldUser {$1=newUser ; $6=newUserHomeDir; } {print $0 ; } ' ${ThePasswordFile} fi Hope this helps. Alan -Original Message- From: kaio [mailto:[EMAIL PROTECTED] Sent: April 4, 2003 10:48 To: [EMAIL PROTECTED] Subject: WinXP username with spaces -- wmaker ko i have my administrator user named with a space on WInXP. cygwin installation is ok, but windowmaker don't install: seems that the space generates a new parameter in installation commands! in fact, wmaker installation creates a new dir in /home/ with the first half of username! -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info:
ar: .libs/libmain.al: Is a directory
Hallo cygwin, getting this strange error here: ar: .libs/libmain.al: Is a directory When I look into the .libs directory I see: $ ls -la .libs total 1 drwxrwxrwx+ 2 gerrit Domänen-0 Apr 5 19:15 ./ drwxr-xr-x+ 4 gerrit Domänen-28672 Apr 5 19:14 ../ -rw-rw-rw-1 gerrit Domänen-8 Apr 5 19:15 libmain.al Any ideas about the problem here? Gerrit -- =^..^= -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: perl 5.8.0 debugger requires double ENTER key to act on commands: cygwin 1.3.22
Gerrit suggests: See the second link above, try to set: $ export PERLIO=perlio in your environment. This also fixes a problem I've had for a long time: I was unable to install Bundle::LWP from CPAN with 5.8.0-2 (many self-tests failed). -- I ... don't want programming to be really interesting any more than I want my toaster to be really interesting. I want my toaster to get the job done and I do not want to hear about exciting new tungesten alloys in the heating wires nor do I want to subscribe to Toaster Today magazine. -- Daniel Weinreb [EMAIL PROTECTED], ll1 mailing list -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: nano and TERM
Hmm, unable to reproduce your problem here so far, prehaps some more details would be useful. Gareth - nano packaging type guy. Thank you for getting back to me. This is what I meant. When Cygwin is installed (i.e. after running setup, filesystems are mounted, registry entries are made, etc, etc) all the supplied applications (such as nano) operate very nicely in the default bash console window and, it turns out, in a rxvt window. I have experimented with the -tn terminal switch available with rxvt but to be honest have found most settings (e.g. -tn cygwin, -tn rxvt, -tn vt100; or no setting at all) to be completely irelevant: things just work. (Though I remember -tn linux turned out to be a doomed selection.) Well done all the packagers. Also, nano and pico and a few other things work seamlessly, and with no obvious differences, in a Windows DOS box. Especially the command-line stuff like ls, find, md5sum, diff, cmp, ... This is because I have put c:\Cygwin\bin in my Windows PATH (early). This offers a huge and valuable extension to the supplied Windows command structure. (I know http://www.nano-editor.org supply a zipped version of nano claimed to be for Windows but since on un-zipping the provision, it turns out to include cygwin1.dll, I deduce it's the same, ot not much different to, the entity maintained by Gareth. So, since nano works in all possible required environments that I can think of (Cygwin-bash, Cygwin-rxvt, Windows-DOS) I have not bothered with this zipped item.) My reason for writing was to report (very unclearly, I now perceive) the following. Occasionally it suits me to umount -A and also change the name c:\Cygwin to c:\SomethingElse. (Don't ask. I'm interested in sparse systems. Particularly in extending the Windows command structure through simply making available the contents of c:\SomethingElse\bin\, and *without* setting up Cygwin in the mounted registry sense.) It seems to me that as long as c:\SomethingElse\bin is in the Windows PATH, then one still has the luxury of Cygwin's applications. Certainly I find am still able to use ls, find, md5sum, cmp, diff, ..., loads of others, ..., and pico. I am more or less certain that I have successfully used nano in these circumstances. But currently I am getting the error message earlier described: Error opening terminal : cygwin. Fergus P.S. I know. This is saying Thank you for supplying an item intended to be used under prevailing circumstances A. I want to use it under prevailing circumstances B, but I find I can't. Kindly mend it so I can. I really am not saying this, and I hope I have no unreasonable expectations. But I guess what I am saying is: (a) I'm sure I could once, and (b) other not dissimilar things seem to work OK under prevailing circumstances B. Is there any possibility that nano could be looked at, because it is SUCH a useful editor. THANK YOU. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: CVS Tags for 1.3.22
Thanks! I downloaded the snapshot and played around with the dates and tags until I had everything matching. BTW, I did use the branch dontuse-21 for the cygwin directory because it matched the snapshot precisely at the time I ran cvs update. This will change as new updates are checked in, of course, and require additional time stamps It would be nice to have official tags but I assume this has been discussed in the past and I don't want to start a new discussion. Thanks, --Christian On Sun, 30 Mar 2003, Igor Pechtchanski wrote: On Sun, 30 Mar 2003 [EMAIL PROTECTED] wrote: Hi, I couldn't find any CVS tags to build Cygwin 1.3.22 from source; the best I could find was the branch dontuse-21 which doesn't seem to be correct. Is there some documentation around this somewhere on the Cygwin web pages? If not, could someone let me know which tags/time stamps to use? I don't want to get the 64-bit file offsets which are being worked on in the trunc version by accident Furthermore, is there some documentation about the tags to use for the support libs such as newlib? Thanks, --Christian Christian, This issue has been raised before. The tags are for cgf's convenience only. There are no guarantees that they would correspond to any particular version of the release. One way of figuring out the CVS timestamp to check out is downloading the corresponding source package and looking at its timestamps... I'm not sure how the timestamps will now interact with the development on separate branches, though. Other ways may be suggested by people here. Igor -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: WinXP username with spaces -- wmaker ko
Alan, The one problem I see immediately with *any* automated solution of this sort is that you can actually *have* two users, e.g., Alan Miles and Alan_Miles, in some situations (in large domains, for example). Neither extra options to /bin/mkpasswd nor scripting will help here -- both will do the wrong thing. The only correct solution is fixing the scripts with broken quoting. I think your solution is fine as a temporary band-aid, but this should be done a) manually, and b) once by everyone who has the problem. I don't think it merits automated scripting. The more bug reports we have, the higher the pressure on package maintainers to fix broken scripts. In fact, now that I think of it, a very useful service to the community would be to maintain a list of scripts with broken quoting (and, optionally, a list of patches to those scripts) [and no, I'm not volunteering, as I'm not affected by this problem at all]. I'd be willing to help out with the patches, though. Igor On Sat, 5 Apr 2003, Alan Miles wrote: Igor, Granted - There are issues with the solution - that is why I said partial solution ( I need to resolve ALL issues) The challenge really has to do with /usr/bin/mkpasswd (which gets called by /etc/postinstall/passwd-grp.sh during the install process) From what I can tell, it is the item that generates the password file, which affects HOME and the home directory setup. Often, users don't have access to change their id's, so there Windows logon name may have spaces in them. We cannot chnage that. Maybe I wasn't too clear in my thoughts - until I have a scripted solution (the code fragment I presented), I manually fix the problem by: 1/ Install CYGWIN (say to c:\cygwin) - I install the icon to my desktop and start menu 2/ Using Windows Wordpad, open the /etc/passwd file (e.g., c:\cygwin\etc\passwd) 3/ Change field 1 (the user name) - change any spaces to _ characters (i.e., change Alan Miles to Alan_Miles) 4/ Change field 6 (the home directory setting) - change any spaces to _ characters (i.e., change /home/Alan Miles to /home/Alan_Miles) 5/ Save /etc/passwd file (e.g., c:\cygwin\etc\passwd) 6/ Click on the cygwin desktop icon to logon to cygwin for the first time. /etc/profile creates the home directory the first time: (fragment from /etc/profile: # If the home directory doesn't exist, create it. if [ ! -d $HOME ]; then mkdir -p $HOME # copy skeleton files cp -a /etc/skel/. $HOME fi ) Now the cygwin creates the home directory properly (i.e., /home/Alan_Miles), HOME is set to /home/Alan_Miles, and USER is set to Alan_Miles I have not had any trouble with this manual setup. This is the item I would like to either fix by scripting, e.g., the partial code fragment already presented, or better still patch /usr/bin/mkpasswd to convert any names it sees with spaces in them to _ characters. Then problem completely solved. I appreciate your input. BTW, Igor got your other message regarding Perl - you said: This wasn't meant to chide you; You weren't chiding me - my first message I thought went to Gerrit only and should have gone to the list (which is why I then resent it). All you did was point out that I inadvertently duplicated messages - no problem. Alan -Original Message- From: Igor Pechtchanski [mailto:[EMAIL PROTECTED] Sent: April 5, 2003 10:53 To: [EMAIL PROTECTED] Cc: kaio; [EMAIL PROTECTED] Subject: RE: WinXP username with spaces -- wmaker ko Alan, The fix below has a few problems (one of which is unconditionally removing /etc/passwd). IMO, it's better to fix the scripts that have improper quoting, or, barring that, unset HOME in your Windows environment (or modify /etc/profile) and make sure your username doesn't contain a space. Going out of your way to accomodate scripts without proper quoting will only lead to the proliferation of such scripts. Igor On Sat, 5 Apr 2003, Alan Miles wrote: All, I have the same problem (on my machine at home) - However I do have a (partial) solution (I am still working on the final solution): This has to be done after CYGWIN installation, but BEFORE and user logon. Edit the /etc/passwd file and replace the space in the first field, and edit the field /home/something with space to /home/something_with_out_space and save. I am working on a scripting solution to install as part of my installation - maybe this could be incorporated as part of the main CYGWIN release. Below is a script fragment on how I am proposing to automate this fix. ThePasswordFile=/etc/passwd ModUserName=$(/usr/bin/echo ${USER} | /usr/bin/sed -e 's/ /_/g') ... ModHomeUserName=$(/usr/bin/basename ${HOME} | /usr/bin/sed -e 's/ /_/g') ModHomeUserDirName=$(/usr/bin/dirname ${HOME})/ ModHome=${ModHomeUserDirName}${ModHomeUserName} /usr/bin/rm -f ${ThePasswordFile} if [ ! -e ${ThePasswordFile} -a ! -L ${ThePasswordFile} ] then
Re: nano and TERM
On Sat, 5 Apr 2003 [EMAIL PROTECTED] wrote: Hmm, unable to reproduce your problem here so far, prehaps some more details would be useful. Gareth - nano packaging type guy. Thank you for getting back to me. This is what I meant. When Cygwin is installed (i.e. after running setup, filesystems are mounted, registry entries are made, etc, etc) all the supplied applications (such as nano) operate very nicely in the default bash console window and, it turns out, in a rxvt window. I have experimented with the -tn terminal switch available with rxvt but to be honest have found most settings (e.g. -tn cygwin, -tn rxvt, -tn vt100; or no setting at all) to be completely irelevant: things just work. (Though I remember -tn linux turned out to be a doomed selection.) Well done all the packagers. Also, nano and pico and a few other things work seamlessly, and with no obvious differences, in a Windows DOS box. Especially the command-line stuff like ls, find, md5sum, diff, cmp, ... This is because I have put c:\Cygwin\bin in my Windows PATH (early). This offers a huge and valuable extension to the supplied Windows command structure. (I know http://www.nano-editor.org supply a zipped version of nano claimed to be for Windows but since on un-zipping the provision, it turns out to include cygwin1.dll, I deduce it's the same, ot not much different to, the entity maintained by Gareth. So, since nano works in all possible required environments that I can think of (Cygwin-bash, Cygwin-rxvt, Windows-DOS) I have not bothered with this zipped item.) My reason for writing was to report (very unclearly, I now perceive) the following. Occasionally it suits me to umount -A and also change the name c:\Cygwin to c:\SomethingElse. (Don't ask. I'm interested in sparse systems. Particularly in extending the Windows command structure through simply making available the contents of c:\SomethingElse\bin\, and *without* setting up Cygwin in the mounted registry sense.) It seems to me that as long as c:\SomethingElse\bin is in the Windows PATH, then one still has the luxury of Cygwin's applications. Certainly I find am still able to use ls, find, md5sum, cmp, diff, ..., loads of others, ..., and pico. I am more or less certain that I have successfully used nano in these circumstances. But currently I am getting the error message earlier described: Error opening terminal : cygwin. Fergus P.S. I know. This is saying Thank you for supplying an item intended to be used under prevailing circumstances A. I want to use it under prevailing circumstances B, but I find I can't. Kindly mend it so I can. I really am not saying this, and I hope I have no unreasonable expectations. But I guess what I am saying is: (a) I'm sure I could once, and (b) other not dissimilar things seem to work OK under prevailing circumstances B. Is there any possibility that nano could be looked at, because it is SUCH a useful editor. THANK YOU. Fergus, Any application depending on ncurses or curses will probably use the terminfo database. For that, it'll need access to /usr/share/terminfo, and thus it'll need at least the / mount. Alternatively, some older applications use /etc/termcap (with the same requirement). I doubt there will be any resolution that will allow you to use a full-screen editor without access to terminfo or termcap. I'm not sure why pico works, possibly because it has an internal termcap entry for a dumb terminal and uses that when the terminfo/termcap database cannot be found. This might have been the case with nano as well, but then its (upstream) maintainers must have decided that terminfo/termcap databases are always available, and that it makes sense to eliminate that functionality. If so, it's unlikely it'll be reinstated. Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: command prompt window title
Randall, You were right. My path was setup so that instead of picking up /usr/bin/vi, it was picking up my DOS version of vi. Once I corrected that, the window title always shows Cygwin and never changes regardless of what I'm running. Thanks, -Anoop PS: Thanks to all who replied. I'm not subscribed to the Cygwin list to it took me a while to check the responses in the archives and get back. === At 11:38 2003-03-31, you wrote: I'm running cygwin's latest version 1.3.22-1. When I run cygwin from my desktop, it brings up a command prompt window with a title Cygwin. I modified cygwin.bat to make my default shell tcsh. Anyway, after I run vi, the window title changes to vi fname where fname is the file that I'm editing. When I quit vi, the title of the window remains that way until I quit the window, or I use vi again with another file, in which case it changes to vi newfilename. It sounds to me like there's a script or other intervening command definition handling your invocation of vi. Please show us the output of where vi (assuming, as you state above, that you're using tcsh). If you're using BASH, then show us the output of type -a vi. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: nano and TERM
On Sat, Apr 05, 2003 at 06:42:44PM +0100, [EMAIL PROTECTED] wrote: (I know http://www.nano-editor.org supply a zipped version of nano claimed to be for Windows but since on un-zipping the provision, it turns out to include cygwin1.dll, Sigh. I've sent them excruciatingly polite email. I expect an indignant response soon. Or maybe I'll be surprised... cgf -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: WinXP username with spaces -- wmaker ko
Igor, I appreciate your input - you can actually *have* two users, e.g., Alan Miles and Alan_Miles, That possibility had not occurred to me - thanks for pointing that flaw out. Since, this for me, is a problem at home, and not at work -I would have to investigate this in my own time. The perl items I need for both home and work, so I can investigate that item there. I could (when I have some time) look at determining what scripts need fixing - once such broken script would be /etc/profile: cp -a /etc/skel/. $HOME should probably read: cp -a /etc/skel/. $HOME Igor, if I could get a list of such patches together, could you provide me with a sanity check (like you already have done). That way any flawed reasoning can be dealt with. Meanwhile, Kaio, if you don't have two users, maybe (at least for the moment), my manual solution would fix your problem. Cheers, Alan -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: Mozilla 1.3 built on cygwin?
If you have such great insight into this type of thing, it won't take you any time at all to duplicate. You've been complaining about this and other cygwin performance issues for months. Why don't *you* do something? I figured fork/exec/signals out from scratch. Certainly the brighter bulbs than I who would be finding the problem would have no problems generating a new and better implementation. It's a wonder why no one has done so yet. There are several people on the list with more skills as I have and I was hoping that somebody of this gurus could fix this I'm wondering too. It seems really it is on me to buy this Nebett book and to see, what I can do. It will need some time to get familiar with this stuff and I have currently two libtool relating outstanding tasks, which has to be finished first You could buy the Nebett book. Or instead download the sample source code (including the fork example) from http://www.newriders.com/content/images/1578701996/downloads/1578701996.zip and http://www.newriders.com/content/images/1578701996/downloads/ntdll.zip. You can find most of the API that's documented in Nebbet's book at http://undocumented.ntinternals.net/. Chris -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: nano and TERM
*** Igor Pechtchanski ([EMAIL PROTECTED]) wrote today: IP I'm not sure why pico works, possibly because it has an internal IP termcap entry for a dumb terminal and uses that when the IP terminfo/termcap database cannot be found. Pico as built for cygwin uses the termcap entries found in /etc/termcap, there's no internal termcap to it. Pico (for cygwin) does not use terminfo. Nor does Pine. My experience as maintaner of Pico (and Pine) is that Pine/Pico have less problems in terms of control of the screen when termcap is used (as opposed to terminfo/curses/ncurses). Pico defaults to some settings (row, colums) when it can not find them, but advanced screen control is done by the termcap library. -- Eduardo http://www.math.washington.edu/~chappa/pine/ -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: 1.3.22: TMPDIR Initialization
Robert and Thorsten, It seems to me the need for the temp folder to be accessible for all users is a global requirement, not for any one particular package. It also seems like a problem you would want to solve once. The worse possible situation would be a number of different packages solving this in incompatible ways. Is there a global/system wide postinstall script? If so, could you point me in the right direction, and I'll take a look at what I think might be a good solution? Thanks in advance. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
Re: WinXP username with spaces -- wmaker ko
| | cp -a /etc/skel/. $HOME | Hi! Just a quick thought, any special reason HOME doesn't contain a legal path, escaped, as it probably (?) should be on a real unix system? (Escaped = /home/Alan\ Miles/ instead of /home/Alan Miles/) What happens if you add the escape-character to /etc/passwd ? /Andy -- The eye of the beholder rests on the beauty! -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: WinXP username with spaces -- wmaker ko
On Sat, 5 Apr 2003, Alan Miles wrote: Igor, I appreciate your input - you can actually *have* two users, e.g., Alan Miles and Alan_Miles, That possibility had not occurred to me - thanks for pointing that flaw out. Since, this for me, is a problem at home, and not at work -I would have to investigate this in my own time. The perl items I need for both home and work, so I can investigate that item there. I could (when I have some time) look at determining what scripts need fixing - once such broken script would be /etc/profile: cp -a /etc/skel/. $HOME should probably read: cp -a /etc/skel/. $HOME Igor, if I could get a list of such patches together, could you provide me with a sanity check (like you already have done). That way any flawed reasoning can be dealt with. [snip] Cheers, Alan Alan, I'm sure if you post the patches (or a link to a website containing them) to this list, you'll get plenty of feedback. I'll certainly take a look at them, although someone may beat me to it. FYI, it's best if the patches are in unidiff (diff -up) format, as per http://cygwin.com/contrib.html (even though your patches won't be to the Cygwin source tree). That way they are easier to review. Igor P.S. If you provide a patch, make sure to specify exactly which package and which version the patch is for... -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: 1.3.22: TMPDIR Initialization
On Sat, 5 Apr 2003, Mark Moore wrote: Robert and Thorsten, It seems to me the need for the temp folder to be accessible for all users is a global requirement, not for any one particular package. It also seems like a problem you would want to solve once. The worse possible situation would be a number of different packages solving this in incompatible ways. Is there a global/system wide postinstall script? If so, could you point me in the right direction, and I'll take a look at what I think might be a good solution? Thanks in advance. Mark, There is no concept of a system wide postinstall script. All unexecuted scripts after an install are run by setup.exe indiscriminately. There are packages that contain *only* postinstall scripts, each of which serves a specific purpose (e.g, base-files, base-passwd, etc). The way it is ensured that the postinstall scripts from these packages are run before some other scripts is by including them into the requires clause of the relevant packages in setup.ini (since the new setup.exe runs postinstall scripts in order of package dependences). If you wish to provide a postinstall script to fix the TMPDIR problem, it would belong either in base-files or in a new package (which you yourself could create and maintain). Igor -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
RE: WinXP username with spaces -- wmaker ko
Igor, Will do. Thanks for your input. Alan -Original Message- From: Igor Pechtchanski [mailto:[EMAIL PROTECTED] Sent: April 5, 2003 14:03 To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: RE: WinXP username with spaces -- wmaker ko On Sat, 5 Apr 2003, Alan Miles wrote: Igor, I appreciate your input - you can actually *have* two users, e.g., Alan Miles and Alan_Miles, That possibility had not occurred to me - thanks for pointing that flaw out. Since, this for me, is a problem at home, and not at work -I would have to investigate this in my own time. The perl items I need for both home and work, so I can investigate that item there. I could (when I have some time) look at determining what scripts need fixing - once such broken script would be /etc/profile: cp -a /etc/skel/. $HOME should probably read: cp -a /etc/skel/. $HOME Igor, if I could get a list of such patches together, could you provide me with a sanity check (like you already have done). That way any flawed reasoning can be dealt with. [snip] Cheers, Alan Alan, I'm sure if you post the patches (or a link to a website containing them) to this list, you'll get plenty of feedback. I'll certainly take a look at them, although someone may beat me to it. FYI, it's best if the patches are in unidiff (diff -up) format, as per http://cygwin.com/contrib.html (even though your patches won't be to the Cygwin source tree). That way they are easier to review. Igor P.S. If you provide a patch, make sure to specify exactly which package and which version the patch is for... -- http://cs.nyu.edu/~pechtcha/ |\ _,,,---,,_[EMAIL PROTECTED] ZZZzz /,`.-'`'-. ;-;;,_[EMAIL PROTECTED] |,4- ) )-,_. ,\ ( `'-' Igor Pechtchanski '---''(_/--' `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-. Meow! Knowledge is an unending adventure at the edge of uncertainty. -- Leto II -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/