Re: LSR lovers: walk the walk
On Sun, Jul 3, 2011 at 7:17 PM, Reinhold Kainhofer wrote: > From a contributors point of view: If a snippet compiles with current stable, > fine, it can be used at LSR. However, if it doesn't compile with current > stable, I can't add it to the LSR now, so it probably won't get added to the > LSR ever. If the release cycle does get shorter (wishful thinking inside ;), I can imagine making do with the current system. Maintaining a "stable-only" LSR does make sense, even if it means commenting out some snippets until they can be handled (see below). > We also have several snippets with workarounds for 2.12, where some > functionality has been added to 2.14 (like the compound time signatures). I > think the LSR should more clearly state the version for which a snippet is > actually intended. Indeed. However, the situation looked much worse to me a couple years ago when waiting for the 2.12 upgrade. It may because I've taken a step back since then, but it does seem less of a bloody mess today. This is also one of the reason why I introduced the notion of tags in lsr. We used to have only one "version-specific" tag that allowed me to browse through snippets that had a chance of being either outdated or fully commented out; we could improve upon this system by having a number of version-specific tags, such as "2.14", "2.15", "2.12" etc. Granted, this wouldn't solve the problem of multiple-version requirement (and possibly the necessity of temporary commenting out some snippets), but it would be a convenient (albeit inelegant) workaround, furthermore one we can set up in 15 seconds right now. > People looking for compound signatures will find that snippet now, although > the > functionality has been added to 2.14. On the other hand, those who have to > stay with 2.12 for some reason are interested in the 2.12-only snippet. I'm gonna go all grahamish on this one: we can't, won't, don't care enough to, support these users. There are a number of people still using 2.10 'cause it came with their distro, 2.8 'cause it's what they're used to, 1.6 'cause they like the TeX backend... (However, here again the tagging ability of the LSR may be of use to some of these people.) Anyways, the "LSR lover" I am has now pushed a minor update onto master. Nothing too extraordinary. Once again, Phil and I are available for any LSR-related question, a snippet that won't be accepted or whatevs. Perhaps the easiest way to go, btw, would be to simply add a basic contact-form to the LSR website, where people can send bleeding-edge/outdated/uncompilable snippets for us to manually review/comment out/translate/convert/commit as we see fit. Cheers, Valentin. ___ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
Re: LSR lovers: walk the walk
Am Sonntag, 3. Juli 2011, 18:54:02 schrieb Graham Percival: > On Sun, Jul 03, 2011 at 06:46:22PM +0200, Reinhold Kainhofer wrote: > > Am Sonntag, 3. Juli 2011, 17:31:19 schrieb Graham Percival: > > > 1. we need to ask Sebastiano to upgrade LSR to 2.14.1. This may > > > require extra manual attention to upgrade some snippets that > > > convert-ly can't handle. > > > > Actually, I have been thinking about implementing multiple-version > > support in LSR (i.e. snippets for 2.12, 2.14 and master). The database > > already has some fields min_version and max_version, but they have not > > been used since 2.4.3... > > I don't think we should ever use unstable for lsr. If a snippet > compiles in the current stable, then just use stable; if it > requires unstable, then I forsee too many people getting confused > about that. >From a contributors point of view: If a snippet compiles with current stable, fine, it can be used at LSR. However, if it doesn't compile with current stable, I can't add it to the LSR now, so it probably won't get added to the LSR ever. We also have several snippets with workarounds for 2.12, where some functionality has been added to 2.14 (like the compound time signatures). I think the LSR should more clearly state the version for which a snippet is actually intended. People looking for compound signatures will find that snippet now, although the functionality has been added to 2.14. On the other hand, those who have to stay with 2.12 for some reason are interested in the 2.12-only snippet. Similarly, many of us are on master even for production systems. Cheers, Reinhold -- -- Reinhold Kainhofer, reinh...@kainhofer.com, http://reinhold.kainhofer.com/ * Financial & Actuarial Math., Vienna Univ. of Technology, Austria * http://www.fam.tuwien.ac.at/, DVR: 0005886 * LilyPond, Music typesetting, http://www.lilypond.org ___ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
Re: LSR lovers: walk the walk
On Sun, Jul 03, 2011 at 06:46:22PM +0200, Reinhold Kainhofer wrote: > Am Sonntag, 3. Juli 2011, 17:31:19 schrieb Graham Percival: > > 1. we need to ask Sebastiano to upgrade LSR to 2.14.1. This may > > require extra manual attention to upgrade some snippets that > > convert-ly can't handle. > > Actually, I have been thinking about implementing multiple-version support in > LSR (i.e. snippets for 2.12, 2.14 and master). The database already has some > fields min_version and max_version, but they have not been used since 2.4.3... I don't think we should ever use unstable for lsr. If a snippet compiles in the current stable, then just use stable; if it requires unstable, then I forsee too many people getting confused about that. Remember, we're looking at a new stable version every 3-4 months. That's not too long to wait for until a new feature is ready. Cheers, - Graham ___ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
Re: LSR lovers: walk the walk
Am Sonntag, 3. Juli 2011, 17:31:19 schrieb Graham Percival: > 1. we need to ask Sebastiano to upgrade LSR to 2.14.1. This may > require extra manual attention to upgrade some snippets that > convert-ly can't handle. Actually, I have been thinking about implementing multiple-version support in LSR (i.e. snippets for 2.12, 2.14 and master). The database already has some fields min_version and max_version, but they have not been used since 2.4.3... > 3. we need somebody to help Reinhold install tomcat. tomcat's not the problem any more ;-) The problem now is to get ERW to work... Cheers, Reinhold -- -- Reinhold Kainhofer, reinh...@kainhofer.com, http://reinhold.kainhofer.com/ * Financial & Actuarial Math., Vienna Univ. of Technology, Austria * http://www.fam.tuwien.ac.at/, DVR: 0005886 * LilyPond, Music typesetting, http://www.lilypond.org ___ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
LSR lovers: walk the walk
Ok, lots of people "talked the talk" about how much they love LSR. Now it's time to "walk the walk". 1. we need to ask Sebastiano to upgrade LSR to 2.14.1. This may require extra manual attention to upgrade some snippets that convert-ly can't handle. 2. we need to add everything in Documentation/snippets/new/ to LSR, and delete them from git. There's 75 snippets. I recommend recruiting people from -user, but it's your responsibility to train and manage them. (for whoever "you" accepts this task) 3. we need somebody to help Reinhold install tomcat. These are not "regular maintenance" tasks, so I'm not aiming this at Phil. You love LSR? be demonstrative. Show it with flowers^H^H^H^H^H^Hoffers of help. (and don't bug me about any of these points, because I'm not one of the LSR-lovers) Cheers, - Graham ___ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel