10749 - StartInstall must return error for null UIName

Status: VERIFIED FIXED

(Core Graveyard :: Installer: XPInstall Engine, defect, P3)

Description

[Daniel Veditz [:dveditz]]

StartInstall must return error for null UIName or else SanityCheck() must not
enforce it not being null -- that's too late.

Comment 1

[Jimmy Lee]

Build 1999-10-20-09-M11(WIN), 1999-10-20-08-M11(MAC), 1999-10-20-08-M11(LINUX)

1. From http://jimbob/trigger2.html, trigger http://jimbob/jars/10749.xpi

var jarSrc = "10749.txt";

b = startInstall(null, "10749", "1.1.1.1", 1);
logComment("startInstall returns = " + b);
f = getFolder("Program");
setPackageFolder(f);
addFile(jarSrc);

RESULT:
Crash.

EXPECTED RESULT:
Error message returned.

NOTE:
Talkback report was sent, but there is still a problem with querying.  Ugh.

Comment 2

[Cathleen]

crash in function nsInstall::AddSubcomponent() at
*aReturn = mVersionInfo->ToString(version);


10-20-99 debug build stack trace:
--------------------------------------
nsInstall::AddSubcomponent(const nsString & {...}, int * 0x032af4f0) line 591 +
25 bytes
InstallAddSubcomponent(JSContext * 0x030df420, JSObject * 0x0270f930, unsigned
int 1, long * 0x02711f64, long * 0x032af644) line 541 + 25 bytes
js_Invoke(JSContext * 0x030df420, unsigned int 1, unsigned int 0) line 672 + 26
bytes
js_Interpret(JSContext * 0x030df420, long * 0x032afef4) line 2247 + 15 bytes
js_Execute(JSContext * 0x030df420, JSObject * 0x0270f930, JSScript * 0x0313d9d0,
JSFunction * 0x00000000, JSStackFrame * 0x00000000, int 0, long * 0x032afef4)
line 845 + 13 bytes
JS_EvaluateUCScriptForPrincipals(JSContext * 0x030df420, JSObject * 0x0270f930,
JSPrincipals * 0x00000000, const unsigned short * 0x025d00a8, unsigned int 646,
const char * 0x00000000, unsigned int 0, long * 0x032afef4) line 2638 + 27 bytes
JS_EvaluateUCScript(JSContext * 0x030df420, JSObject * 0x0270f930, const
unsigned short * 0x025d00a8, unsigned int 646, const char * 0x00000000, unsigned
int 0, long * 0x032afef4) line 2620 + 35 bytes
JS_EvaluateScript(JSContext * 0x030df420, JSObject * 0x0270f930, const char *
0x030df590, unsigned int 646, const char * 0x00000000, unsigned int 0, long *
0x032afef4) line 2587 + 33 bytes
RunInstallOnThread(void * 0x030dfbc0) line 358 + 30 bytes
_PR_NativeRunThread(void * 0x030df950) line 391 + 13 bytes
_threadstartex(void * 0x030df850) line 212 + 13 bytes
KERNEL32! 77f04f2c()

Comment 3

[leger]

Clearing FIXED resolution due to reopen.

Comment 4

[Daniel Veditz [:dveditz]]

This bug is fixed. If the crash is reproduceable it's a new bug (likely the
same that prevented verification of the Multi-trigger bug). The crash may block
verification of this bug, but there is no evidence of a regression.

Comment 5

[Jimmy Lee]

Build 1999-11-01-09-M11(WIN), 1999-11-01-08-M11(MAC), 1999-11-01-08-M11(LINUX)

There still is a crash.  This crash which Dan claims may be different is
documented from Bug 17774.  Marking this one Verified.

Comment 6

[leger]

Bulk move of XPInstall (component to be deleted) bugs to Installer: XPInstall
Engine

Comment 7

[David Epstein]

Moved test xpi (10749.xpi) from jimbob/jars to jimbob/bugs folder.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/9c613f10c2b7
[components] For https://github.com/mozilla-mobile/android-components/issues/10749 - New NonBlockingHttpIconLoader