Hi All, After having conversations with Cloph and shm_get I came up with the following list of bugs that should *have something related to bibisect*. What this means is that one of these tags should exist:
_*Whiteboard*_ 1) bibisectRequest; 2) bibisected; 3) notBibisectable (this means it falls in the appropriate range but for some reason it was not bibisectable...usually a series of bad commits that make the range useless); 4) preBibisect (in which case the version field should be updated to reflect this - *applies to bugs depending on versions listed below*). _*Keyword*_ bisected - this applies *only* when the _exact_ commit is determined - this isn't done through the normal bibisect method, it requires extra steps. After doing this IMHO we should still include "bibisected" in the whiteboard status. *_Versions Covered:_ *Linux - any bug after 3.5beta0 that is a regression (including master); Windows - any bug 4.3 - master (inclusive); Mac - 4.1 - 5.0 (inclusive - note there is a gap in the 4.3 series....if you come across one of these the correct tag is "notBibisectable" because of the gap) _*Useful Links *_These links are bugs that I believe match the above criteria for each platform - so in theory *each of these bugs should have some tag pertaining to bibisect **Linux: *http://tinyurl.com/linuxBibisect* **Windows: *http://tinyurl.com/windowsBibisect *Mac: *http://tinyurl.com/macBibisect _*Regression Tagging *_Moving forward it's really important that we continue to improve tagging regressions (we have indeed gotten a lot better!). So I've found a couple things to help: 1) Have 3.3 installed and at least check current fresh and 3.3 (if it's in 3.3, we know not a regression, can update version) 2) Ask the user to test their own bug - link to the page with old releases (http://downloadarchive.documentfoundation.org/libreoffice/old/) - say it helps a lot and it _could_ potentially help get the bug fixed faster; 3) Quickly explain what version field means (oldest version not latest version) There are several hundred bugs available - and this is an under-estimate as they only include bugs that are tagged as "regression" - if I had to guess we have _at minimum_ twice as many bugs that could use a bibisect. Correct tagging is crucial moving forward :-D Best, Joel
_______________________________________________ List Name: Libreoffice-qa mailing list Mail address: Libreoffice-qa@lists.freedesktop.org Change settings: http://lists.freedesktop.org/mailman/listinfo/libreoffice-qa Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette List archive: http://lists.freedesktop.org/archives/libreoffice-qa/