The regression happens only when trying to close using <ctrl>+<q> or <alt>+<f4>. The save pop-up shows properly trying to use File+Close from the File menu.
I tried to bisect with VCL=kf5 but it works perfectly showing always the save pop-up as expected.
The regression is gtk3 only.
For my bisect I was able to reproduce forcing VCL with:
SAL_USE_VCLPLUGIN=gtk3
cd8a95897491fa33980122053825176779e7624a is the first bad commit
commit cd8a95897491fa33980122053825176779e7624a
Author: Jenkins Build User <email address hidden>
Date: Fri Feb 26 01:04:33 2021 +0100
Some notes for reproducing the regression:
The regression happens only when trying to close using <ctrl>+<q> or <alt>+<f4>. The save pop-up shows properly trying to use File+Close from the File menu.
I tried to bisect with VCL=kf5 but it works perfectly showing always the save pop-up as expected.
The regression is gtk3 only. VCLPLUGIN= gtk3
For my bisect I was able to reproduce forcing VCL with:
SAL_USE_
cd8a95897491fa3 398012205382517 6779e7624a is the first bad commit 398012205382517 6779e7624a
commit cd8a95897491fa3
Author: Jenkins Build User <email address hidden>
Date: Fri Feb 26 01:04:33 2021 +0100
source sha:f7243a465fc 75f8c29c1decb76 28f7a8f6d1d769
source sha:f7243a465fc 75f8c29c1decb76 28f7a8f6d1d769
instdir/ program/ libsmlo. so | Bin 1791192 -> 1791264 bytes program/ setuprc | 2 +- program/ versionrc | 2 +-
instdir/
instdir/
3 files changed, 2 insertions(+), 2 deletions(-)
git bisect log ed8281d59f22bcf bdd13119279] source sha:91385bbc7cc cfdf59f60a24eaf 81894772134af0 e51d58dff4d0b4c e9022013411] source sha:574c5709064 2347980d2395e1e 183cc7b5c171ad d81eba53ed8d55c 23543d82ef7] source sha:487df0c8f30 7500029b06c8998 5a1148ebc784b5 81eba53ed8d55c2 3543d82ef7 e1013e2ea8f511f 9ace5f6da4b] source sha:ebea072fa7d 7081ba9d1e549f5 aa629c749d8c0e 1013e2ea8f511f9 ace5f6da4b bd983a21610e6a7 45470c4b82d] source sha:14fb7d2cce5 c41bacf7c6b571a a946c766205bc3 d983a21610e6a74 5470c4b82d fd388e43c6fb266 16a8d542ccb] source sha:bc0d161d365 1ab2d38a6ff664e 43841d2719f4cc d388e43c6fb2661 6a8d542ccb 336b9b78008ce1b d9239a092ab] source sha:608151c7de0 f0661783daceecd 9250bdcf497b59 36b9b78008ce1bd 9239a092ab 29322817145ecc6 fd48798ccfd] source sha:11d344f6271 e4171f8006c2cb2 9372f612cbdc00 9322817145ecc6f d48798ccfd 7099e00b069e2f4 5f2d0c532ea] source sha:ff94f777923 2088728bf724ec4 7608f35f5310b7 099e00b069e2f45 f2d0c532ea 2410eb65d6b3c21 dc52590187d] source sha:d92235df758 29a8cf2ee8cc7b0 b76063093b6cc2 410eb65d6b3c21d c52590187d f3c113adbec19d5 59fc0d25351] source sha:bbafee3424f dc59d2f69204fef ab8a318aeec0e1 3c113adbec19d55 9fc0d25351 65fcbc4fd9d86b6 0152511d19c] source sha:f580ea54bdd ee91e85ef018be2 86b60b48935681 5fcbc4fd9d86b60 152511d19c 12d45f412a8eb64 12259dbb107] source sha:1220b206c7b e018ca970421f36 b8ed4774321ae2 2d45f412a8eb641 2259dbb107 339801220538251 76779e7624a] source sha:f7243a465fc 75f8c29c1decb76 28f7a8f6d1d769 398012205382517 6779e7624a ce9db5e23d85a20 6f69711f21a] source sha:2465b0467d5 7ca325e820f13ef eb40e69386b4a1 e9db5e23d85a206 f69711f21a 339801220538251 76779e7624a] source sha:f7243a465fc 75f8c29c1decb76 28f7a8f6d1d769
# bad: [a0d29a846f1429
# good: [36741205b2e1c9
git bisect start 'origin/master' 'oldest'
# good: [a41af4542f3c9b
git bisect good a41af4542f3c9bd
# good: [fdbbff390d14ec
git bisect good fdbbff390d14ece
# good: [71fc1eec4e0ecd
git bisect good 71fc1eec4e0ecdb
# good: [b01d2c96f3a1cf
git bisect good b01d2c96f3a1cff
# good: [17cc6f8c9ce4df
git bisect good 17cc6f8c9ce4df3
# bad: [606a1b6c614931
git bisect bad 606a1b6c6149312
# bad: [e816f0bcf20015
git bisect bad e816f0bcf200157
# bad: [377f98e9d6b643
git bisect bad 377f98e9d6b6432
# bad: [924927e1950ae5
git bisect bad 924927e1950ae5f
# bad: [e615b0b9af48f0
git bisect bad e615b0b9af48f06
# good: [eb71bd63e4f772
git bisect good eb71bd63e4f7721
# bad: [cd8a95897491fa
git bisect bad cd8a95897491fa3
# good: [4933965c83355a
git bisect good 4933965c83355ac
# first bad commit: [cd8a95897491fa
https:/ /git.libreoffic e.org/core/ +/f7243a465fc75 f8c29c1decb7628 f7a8f6d1d769
Adding to CC Caolán McNamara
Raising the priority to high-major because there's a dataloss risk connected to this regression.