Re: [Rpm-maint] [rpm-software-management/rpm] Support libgrypt as crypto library (#826)
@pmatilai Perhaps we may want to add a commit to deprecate NSS backend or just straight rip it out? -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/826#issuecomment-529617826___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
Merged #827 into master. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/827#event-2618729816___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
Works for me. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/827#issuecomment-529480192___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
>>Or just leave them as historical reference for future generations to gape >> at... > I like this option. :) Changed. I leave the original names. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/827#issuecomment-529469822___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
@pavlinamv pushed 1 commit. 8a1f9ec0f420ba4a5eec64db52c5d86ff6162c03 Correct and update Query formats documentation -- You are receiving this because you are subscribed to this thread. View it on GitHub: https://github.com/rpm-software-management/rpm/pull/827/files/209faf3a124341c3e26ba811759a09ec89490603..8a1f9ec0f420ba4a5eec64db52c5d86ff6162c03 ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Rip the marker support for multiline %{expr:...} error messages (#828)
Merged #828 into master. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/828#event-2618623253___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
> Or just leave them as historical reference for future generations to gape > at... I like this option. :) -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/827#issuecomment-529461999___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
pmatilai commented on this pull request. > To make this work properly, you need to tell RPM to always print the first item in the NAME element. You do this by placing a '=' before the tag name, like this: \verbatim -rpm -q --queryformat "[%{=NAME} %{FILENAMES}\n]" cdp +rpm -q --queryformat "[%{=NAME} %{FILENAMES}\n]" systemd If you change this, then you also need to change the output sample below. We can't list the entire systemd package contents there so dunno if it's really worth it. Looking closer, there are several other now extinct package names mentioned in the document too, IF we want to update then maybe we should update them all. Or just leave them as historical reference for future generations to gape at... -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/827#pullrequestreview-285487935___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Rip the marker support for multiline %{expr:...} error messages (#828)
Corrected according to the review. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/828#issuecomment-52945___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Rip the marker support for multiline %{expr:...} error messages (#828)
@pavlinamv pushed 1 commit. 218633033acea714edbbdb6b21a8dba8aa3d39e8 Disable marker on multiline expression error messages -- You are receiving this because you are subscribed to this thread. View it on GitHub: https://github.com/rpm-software-management/rpm/pull/828/files/b2260fbae4087f55615219376906d066db7341ee..218633033acea714edbbdb6b21a8dba8aa3d39e8 ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Improve description of conditionals + comments in spec documentation (#830)
pmatilai commented on this pull request. > +%elifarch, %elifos or %else can be optionally used. Conditionals %endif and +%else should not be followed by any text. Conditionals may be nested within +other conditionals. + +Conditionals are not macros. It has an interesting consequence (similar as +the fact that macros get expanded in comments). If a conditional is a part +of a definition like +\verbatim + %define macro_with_if \ + %if 0 \ + %global output XXX\ + %endif +\endverbatim +the expansion (including %global) occurs before evaluating of the +conditionals. Thus after the expansion of %macro_with_if macro %output +is defined to XXX. I don't think this last block belongs here, at least in this detail. Dark corners are often best left undocumented to allow changing if a saner behavior becomes feasible later. I'd condense that into something like: %if-conditionals are not macros, and are unlikely to yield expected results if used in them. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/830#pullrequestreview-285483065___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Improve description of conditionals + comments in spec documentation (#830)
pmatilai commented on this pull request. > + +%if can be used for various purposes. The test can be evaluated based on +the existence of a macro, like: +\verbatim + %if %{defined with_foo} && %{undefined with_bar} +\endverbatim +string comparison: +\verbatim + %if %{optimize_flags} != "none" +\endverbatim +or a mathematical statement: +\verbatim + %if 0%{?fedora} > 10 || 0%{?rhel} > 7 +\endverbatim +Generally, a mathematical statement allows to use logical operators +&&, ||, !, relational operators !=, ==, <, > , <=, >= and parentheses. Might as well document the rest of the mathematical operators as well (+-/*) -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/830#pullrequestreview-285478700___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Improve description of conditionals + comments in spec documentation (#830)
pmatilai commented on this pull request. > %endif \endverbatim -Conditionals are not macros, thus they can be used in spec files only. +%ifos is used to control RPM's spec file processing according to the +build system's operating system. + +%if can be used for various purposes. The test can be evaluated based on +the existence of a macro, like: +\verbatim + %if %{defined with_foo} && %{undefined with_bar} +\endverbatim +string comparison: +\verbatim + %if %{optimize_flags} != "none" To set a good example, for string comparisons it's best to quote both operands. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/830#pullrequestreview-285477252___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Improve description of conditionals + comments in spec documentation (#830)
pmatilai commented on this pull request. > %endif \endverbatim -Conditionals are not macros, thus they can be used in spec files only. +%ifos is used to control RPM's spec file processing according to the +build system's operating system. It's not build system's operating system, but the build *target* operating system. In > 99% of the cases that's the same thing, but in theory rpm supports cross-architecture and -os builds. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/830#pullrequestreview-285475964___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Improve description of conditionals + comments in spec documentation (#830)
pmatilai commented on this pull request. > \verbatim - %ifarch ppc64 ppc x86_64 - ... - %if 0%{?ver} && 0%{?ver} >= 100 - ... - %else - ... - %endif + %ifarch %{power64} s390x x86_64 ia64 + %global architectures 1 The new example is just as meaningles as the old one. While we're at it, lets make the example actually meaningful. As in, something that actually explains why somebody would want to use it, such as this snippet from anaconda.spec: ``` %ifarch s390 s390x BuildRequires: s390utils-devel %endif ``` -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/830#pullrequestreview-285474685___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
Moreover to the described corrections (according to the review) I add an example of example of a query expression. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/827#issuecomment-529451544___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Improve description of conditionals + comments in spec documentation (#830)
pmatilai commented on this pull request. > @@ -214,24 +214,67 @@ can express this as BuildConflicts: gcc <= 2.7.2.1 \endverbatim +\section comments Comments + +Comments in spec file have # at the start of the line. +\verbatim + # this is a comment +\endverbatim +Macros are expanded in comments, thus if a macro is in comment use two percent +characters before it: I'd make that more like: ``` Macros are expanded even in comment lines. If this is undesireable, escape the macro with an extra percent sign (%): <...example..> ``` Now that we have such a thing, it'd be good to mention %dnl as a less quirky way of adding comments to specs. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/830#pullrequestreview-285472092___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
@pavlinamv pushed 1 commit. 209faf3a124341c3e26ba811759a09ec89490603 Correct and update Query formats documentation -- You are receiving this because you are subscribed to this thread. View it on GitHub: https://github.com/rpm-software-management/rpm/pull/827/files/f7c7f469966081328932a6320c22e54643e59952..209faf3a124341c3e26ba811759a09ec89490603 ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Rip the marker support for multiline %{expr:...} error messages (#828)
pmatilai commented on this pull request. > @@ -110,6 +110,12 @@ typedef struct _parseState { static void exprErr(const struct _parseState *state, const char *msg, const char *p) { +const char *s = state->str; + +s = strchr(s,'\n'); +if (s && (*(s+1) != '\0')) + p = NULL; + While "s" is perfectly acceptable variable name in a case like this, a more descriptive one would make the code that little bit more readable. Say, "nl" or even "newline", after which the whole thing reads almost as human language. *Almost* :) -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/828#pullrequestreview-285450831___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Rip the marker support for multiline %{expr:...} error messages (#828)
pmatilai commented on this pull request. > @@ -110,6 +110,12 @@ typedef struct _parseState { static void exprErr(const struct _parseState *state, const char *msg, const char *p) { +const char *s = state->str; + +s = strchr(s,'\n'); Waste of perfectly good screen estate... Why not just ```const char *s = strchr(state->str, '\n');``` -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/828#pullrequestreview-285450509___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Correct and update Query formats documentation (#827)
pmatilai requested changes on this pull request. A welcome update, but lets make it really count. Please address the individual comments. > @@ -136,6 +135,8 @@ readable format in SI resp IEC 8 standard. humansi uses 1K = 1000, 1M = 100, ... humaniec uses 1K = 1024, 1M = 1048576, ... +The formatting tags are listed in lib/formats.c in structure headerFmt_s. Source code is not user documentation and should not be linked as such. The formats are documented in rpm (8) manual in the QUERY OPTIONS section. > @@ -153,10 +154,10 @@ braces. \section queryformat_example Example: Viewing the Verify Flags -The following example query is run against dev because I know %verify -is used there. +The following example query is run against nss because %verify is used +there. I agree the old sentence is pretty bad, but this is not much of an improvement. "nss" doesn't go by that name everywhere, the packaging details such as the name and use of %verify are highly distro specific. These days, "systemd" might well be the most ubiquitous name there is, but that doesn't make %verify flags any less distro specific. > \verbatim - rpm -q --qf '[%{filenames} %{fileverifyflags}\n]' dev + rpm -q --qf '[%{FILENAMES} %{FILEVERIFYFLAGS:hex}\n]' nss To really improve this example, I'd suggest using something else than FILEVERIFYFLAGS for the data. FILEMODES:octal might be a better example as all packages with files have them, and the values are commonly known, so you could delete the rpmfiles.h reference and values from the document too. > @@ -90,11 +89,11 @@ grep or awk). If you try the obvious, rpm -q --queryformat "[%{NAME} %{FILENAMES}\n]" cdp While we're updating package names, here's another candidate. Even I don't remember what "cdp" used to do :D -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/827#pullrequestreview-285425915___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
Re: [Rpm-maint] [rpm-software-management/rpm] Support libgrypt as crypto library (#826)
In a perfect world I'd be much happier with less choice over crypto implementations, but as it is, with our default being something of a cross between a gorilla and an elephant on our back, and the nicer one being license-restricted... Did a quick performance comparison between the variants, ```time ./rpm -qa|wc -l``` on my laptop, results are fastest from three runs each: $time ./rpm -qa|wc -l * openssl: real0m1.609s user0m1.523s sys 0m0.111s * libgcrypt: real0m1.983s user0m1.902s sys 0m0.130s * nss: real0m2.683s user0m2.580s sys 0m0.124s * beecrypt: real0m4.706s user0m4.610s sys 0m0.115s This comparison is quite favorable too: $ wc -l rpmio/digest_*.c 507 rpmio/digest_beecrypt.c 404 rpmio/digest_libgcrypt.c 532 rpmio/digest_nss.c 833 rpmio/digest_openssl.c It's the simplest of them all code-wise and second fastest overall, beating our current default at both and with a nice license. Not to mention just a fraction of the size of the nss on the library side. What's not to like? I'll be happy to merge, but I'd actually suggest going one step further: add a separate commit to make libgcrypt the default. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/826#issuecomment-529410611___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint
[Rpm-maint] [rpm-software-management/rpm] Suppress inhibition lock warning message in single-user mode (#831)
The inhibition lock warning message is output in single-user mode due to absent of dbus service. I suppress the warning message because the inhibit operation may not be needed in this mode and the message just confuses users. You can view, comment on, or merge this pull request online at: https://github.com/rpm-software-management/rpm/pull/831 -- Commit Summary -- * Suppress inhibition lock warning message in single-user mode -- File Changes -- M plugins/systemd_inhibit.c (4) -- Patch Links -- https://github.com/rpm-software-management/rpm/pull/831.patch https://github.com/rpm-software-management/rpm/pull/831.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/rpm-software-management/rpm/pull/831 ___ Rpm-maint mailing list Rpm-maint@lists.rpm.org http://lists.rpm.org/mailman/listinfo/rpm-maint