Bug 189829
| Summary: | New Port: games/tesseract | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Daniel C. Dowse <operator> | ||||||||||||||
| Component: | Individual Port(s) | Assignee: | Jan Beich <jbeich> | ||||||||||||||
| Status: | Closed FIXED | ||||||||||||||||
| Severity: | Affects Only Me | CC: | amdmi3, bf, jbeich, marino, miwi | ||||||||||||||
| Priority: | Normal | ||||||||||||||||
| Version: | Latest | ||||||||||||||||
| Hardware: | Any | ||||||||||||||||
| OS: | Any | ||||||||||||||||
| Attachments: |
|
||||||||||||||||
|
Description
Daniel C. Dowse
2014-05-15 07:50:00 UTC
this is not a good name for the port, there is already graphics/tesseract. Maybe "games/tesseract-fps" ? Without a change there is probably a PKGNAME clash too.
1) NOPORTDATA is probably deprecated (I'm not sure why the data just isn't installed unconditionally)
2) don't have both do-install and post-install targets, just combine them
3) I don't get a one element loop in post-install. Just remove the loop and replace ${f} with {STAGEDIR}${TESS_BIN}
4) we'd like to see poudriere testlogs (testport or bulk -t), or at least redports, tinderbox, or "make check-plist ; make stage-qa"
Created attachment 152637 [details]
shar, improved
I wanted to only review the port but...
- unbundle net/libenet
- respect CXXFLAGS
- depend on SDL2 deps, not SDL 1.2.x
- rewrite conditionals to use option helpers
- compress wrapper scripts into one
- actually write to ~/.tesseract
- provide .desktop file
- replace NOPORTDATA with DATA option
- use DISTVERSION like games/cube
- fix portlint warnings/errors
- drop libGLU and add libX11 dep per |readelf -d|
- drop implicit USE_SDL=sdl
- drop pointless pkg-message and patch-src__Makefile
- reword COMMENT, probably not better
- style: wrap lines, add parens, s/BUILD/FLAVOR/
Created attachment 152638 [details]
improved log for 8.4R amd64
Created attachment 152639 [details]
improved log for 10.1R i386
Created attachment 152640 [details]
shar, improved
Oops, forgot the icon for .desktop file.
$ portlint -AC
looks fine.
(In reply to John Marino from comment #1) > this is not a good name for the port, there is already > graphics/tesseract. Maybe "games/tesseract-fps" ? Without a > change there is probably a PKGNAME clash too. Maybe graphics/tesseract should be renamed to match its distfile. Keeping vendor-given name intact is more consistent with how other ports behave in general. Debian/Ubuntu also use "tesseract-ocr" but otherwise "tesseract" is more popular. Brendan, what do you think? Created attachment 152913 [details] shar, improved, v1.1 Brendan Fabeny <bf1783@gmail.com> writes: > I can't log on to bugzilla at the moment, so I'm replying directly. > If it would make adding this port easier, I've no objection to > changing graphics/tesseract to graphics/tesseract-ocr. > > Regards, > b. Adding PKGNAMESUFFIX here is probably easier. Too early to abandon given the engine is still maintained upstream but no new release yet. The shar(s) here need to be rebased according to ports' latest style. Reset assignee that failed to provide feedback for the past year. A commit references this bug: Author: jbeich Date: Tue Mar 15 16:27:43 UTC 2016 New revision: 411185 URL: https://svnweb.freebsd.org/changeset/ports/411185 Log: games/tesseract{,-data}: add new port PR: 189829 (based on) Submitted by: operator@myrechenzentrum.net Reviewed by: marino (partially, earlier version) Tesseract is a first-person shooter game focused on instagib deathmatch and capture-the-flag gameplay as well as cooperative in-game map editing. http://www.tesseract.gg Changes: head/games/Makefile head/games/tesseract/ head/games/tesseract/Makefile head/games/tesseract/distinfo head/games/tesseract/files/ head/games/tesseract/files/tess_wrapper.in head/games/tesseract/pkg-descr head/games/tesseract-data/ head/games/tesseract-data/Makefile head/games/tesseract-data/distinfo head/games/tesseract-data/pkg-descr Thanks. Committed after fixing PCH issue and turning DATA option into a separate port. Jan, I'm getting this in DragonFly's poudriere for tesseract-data port: =========================================================================== ====> Running Q/A tests (stage-qa) ====> Checking for pkg-plist issues (check-plist) ===> Parsing plist ===> Checking for items in STAGEDIR missing from pkg-plist Error: Orphaned: @dir %%DATADIR%%/media/prefab Error: Orphaned: @dir %%DATADIR%%/screenshot ===> Checking for items in pkg-plist which are not in STAGEDIR ===> Error: Plist issues found. *** Error code 1 Stop. make: stopped in /usr/ports/games/tesseract-data => Error: check-plist failures detected => Cleaning up wrkdir Can you confirm the same on FreeBSD? While splitting attachment 152637 [details] into 2 ports the following lines were lost
XDATA_PLIST_FILES="@dir %%DATADIR%%/media/prefab" \
X "@dir %%DATADIR%%/screenshot"
Before commit I've accidentally tested with |poudriere testport| which builds dependencies *without* QA checks unlike |poudriere bulk -t|.
A commit references this bug: Author: jbeich Date: Sat Mar 26 13:53:47 UTC 2016 New revision: 411926 URL: https://svnweb.freebsd.org/changeset/ports/411926 Log: games/tesseract-data: pacify check-plist ===> Checking for items in STAGEDIR missing from pkg-plist Error: Orphaned: @dir %%DATADIR%%/media/prefab Error: Orphaned: @dir %%DATADIR%%/screenshot PR: 189829 Pointy hat: jbeich (last minute regression) Reported by: marino Changes: head/games/tesseract-data/Makefile PORTDATA is (arguably) broken with empty dirs in hierarchy as it forces ports either to list them via PLIST_DIRS or manipulate TMPPLIST directly. Very few ports actually require empty directories under maybe read-only locations. games/tesseract-data isn't one of such but @dir lines are unlikely to change on updates. Also, PORTTESTING_FATAL="no" in poudriere.conf may help to avoid blocking *package* build due to harmless plist errors scattered in the tree. There's a wide belief that PORTDATA is evil and should be removed from the ports tree completely. I'm sure it's used correctly here but in general it's abused. Couldn't empty directories be purged beforehand though to workaround the issue? I've used this technique myself on a couple of ports where I rolled-my-own plist generator. (In reply to John Marino from comment #16) > There's a wide belief that PORTDATA is evil... >1 Mb plists or ugly TMPPLIST expressions work better a strawman. PORTDATA is no worse than PORTDOCS or PORTEXAMPLES - each behaves as a limited automatic plist. However, PORTDATA can be used in a more generic way by redefining DATADIR. > Couldn't empty directories be purged beforehand though to workaround the issue? If everyone invents their own workaround we have a mess which may scare away new contributors. gecko@ ports are an example to avoid. ;) I've filed bug 208307 just in case. if the left hand represents the good that PORTSDATA does, and the right hand represents the abuse, the contents of the right hand far outweigh the left. Count me among those that would vote to remove it. It's essentially an auto-plist generator so if anything, one of those should be introduced (and used in a *very* limited capacity). or better yet, define a mechanism to skip a PLIST completely, b/c that's the same thing really. |
