507513 - Need ABI defined for WinCE / WinMo

Status: RESOLVED FIXED

(Firefox Build System :: General, defect)

Description

[Robert Strong (they/them - no direct email)]

From readin others bugs it appears that no one has a good handle on what the ABI should be... me included. Even so, we still need this for app update and extension manager

Comment 1

[Robert Strong (they/them - no direct email)]

WINCE_arm-msvc?

Comment 2

[Robert Strong (they/them - no direct email)]

I think this should block

Comment 3

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Agree; essentially blocks update.

WINCE_arm-msvc sounds fine to me.  "arm" is a bit hairy since there are a bunch of variants there, but people with binary components for arm can do runtime detection.

Comment 4

[Robert Strong (they/them - no direct email)]

hmmm... perhaps the following would be better?
WINCE_armv4-msvc
WINCE_armv4i-msvc
WINCE_armv4t-msvc
WINCE_armv5-msvc
etc.?

Since we don't support thumb currently we would require support for WINCE_armv6-msvc for now unless I'm mistaken.

Comment 5

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

No; I think "arm" is actually better, because there's really no sane way to describe the full instruction set.  The version isn't enough; it might or might not have VFP, might or might not have NEON, might or might not have WMMX, etc.  We know that we basically won't support anything below armv5, so that gives developers a base instruction set to target, and they can do runtime detection and use more specific code if appropriate.

Comment 6

[Robert Strong (they/them - no direct email)]

Attachment: patch rev1

This should be all that is necessary. I'm going to compile / verify this before requesting review.

Comment 7

[Robert Strong (they/them - no direct email)]

Comment on attachment 391826 [details] [diff] [review]
patch rev1

Actually, both config.status files have s%@TARGET_XPCOM_ABI@%arm-msvc%g which is what we want so requesting review

Comment 8

[(not currently active) Ted Mielczarek]

We should probably verify that this doesn't make us set an XPCOM ABI for Maemo builds, since AFAIK nobody has said that's defined yet.

Comment 9

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

I hope not, because Maemo isn't using msvc :)

(Note: on maemo, we need to make sure to include 'eabi' in the abi string... this is the new ARM caling convention, linux can use both, but all new-world stuff uses eabi.  So should be something like arm-eabi-gcc4)

Comment 10

[Robert Strong (they/them - no direct email)]

If TARGET_COMPILER_ABI is defined then with this change the ABI will be defined for Maemo. I'll check if it is by sending it to the try server.

Comment 11

[Robert Strong (they/them - no direct email)]

frig... can't do an xpcshell test to check this on the try server since xpcshell doesn't have the ABI set.

Comment 12

[Robert Strong (they/them - no direct email)]

Attachment: patch rev2

Vlad, do we need eabi for WinCE as well?

Comment 13

[Robert Strong (they/them - no direct email)]

Comment on attachment 391928 [details] [diff] [review]
patch rev2

per Vlad arm-msvc shouldn't specify eabi

Comment 14

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

pushed http://hg.mozilla.org/mozilla-central/rev/c7538abbd5a4

Comment 15

[Robert Strong (they/them - no direct email)]

Thanks Paul

Comment 16

[(dormant account)]

Comment on attachment 391928 [details] [diff] [review]
patch rev2

>+arm)
>+    if test "$OS_TARGET" == "WINCE"; then
>+        CPU_ARCH="$OS_TEST"
>+    fi
>+    ;;
> esac

This part doesn't appear to work. should be '='

Comment 17

[Robert Strong (they/them - no direct email)]

I verified earlier that it did work. What specifically are you seeing that makes you think it doesn't work?

Comment 18

[Benjamin Smedberg]

test == is not portable. test = is

Comment 19

[Robert Strong (they/them - no direct email)]

Attachment: followup patch (pushed to mozilla-central)

Comment 20

[(not currently active) Ted Mielczarek]

Comment on attachment 394963 [details] [diff] [review]
followup patch (pushed to mozilla-central)

*sigh*, sorry, I missed that. I hate shellscript.

Comment 21

[Robert Strong (they/them - no direct email)]

Comment on attachment 394963 [details] [diff] [review]
followup patch (pushed to mozilla-central)

Pushed to mozilla-central
http://hg.mozilla.org/mozilla-central/rev/aef07655f47f

Comment 22

[Robert Strong (they/them - no direct email)]

Comment on attachment 394963 [details] [diff] [review]
followup patch (pushed to mozilla-central)

Drivers, need this followup fix for portability on mozilla-1.9.2

Comment 23

[Robert Strong (they/them - no direct email)]

Comment on attachment 394963 [details] [diff] [review]
followup patch (pushed to mozilla-central)

missed that this already has blocking

Comment 24

[Robert Strong (they/them - no direct email)]

followup pushed to mozilla-1.9.2
http://hg.mozilla.org/releases/mozilla-1.9.2/rev/4a152dd44217

Comment 25

[Serge Gautherie (:sgautherie)]

Attachment: (Cv1-CC) Copy it to comm-central [Checkin: Comment 26]

Comment 26

[Serge Gautherie (:sgautherie)]

Comment on attachment 428383 [details] [diff] [review]
(Cv1-CC) Copy it to comm-central
[Checkin: Comment 26]


http://hg.mozilla.org/comm-central/rev/1269773ba53e