Closed Bug 554129 Opened 16 years ago Closed 3 years ago

secutil functions not exported

Categories

(NSS :: Libraries, defect, P5)

x86
Linux

Tracking

(Not tracked)

RESOLVED WONTFIX

People

(Reporter: jdennis, Unassigned)

Details

User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.8) Gecko/20100216 Fedora/3.5.8-1.fc12 Firefox/3.5.8 Build Identifier: 3.12.6 There are a number of useful utilities in secutil which applications would like to use. The very most basic being the ability to pretty print a certificate or to retrieve error strings. Most (but probably not all) of the functions in secutil.h should be in an exported shared library. Yes, I realize one can "roll your own" from the same low level NSS functions the secutil library uses, but why? Applications really shouldn't be concerned if a new certificate extension is added with the need to backport (e.g. copy from secutil.c) the relevant code into their application. Reproducible: Always Steps to Reproduce: Try to print a certificate from an application using NSS. Actual Results: nothing, you can't do it.
The reason that these functions have never been exported is that there has been a long standing requirement from certain corporate management that any exported functions for similar purposes be fully internationalized and localized. The scheme that one company uses for I18n and L10n is not the same as another. What works in Mozilla products may not work in Linux products, or Sun products or ... The lack of a single universal standard API for L10n and I18n is the holdup. If you create a shared library whose functions are not Internationalized and Localized, you will invite a whole raft of bug reports, all complaining about it. There is no single solution that will satisfy all complainers. So, for the last 14 years, NSS's answer has been: we don't do UI.
Thank you for the explanation Nelson, although I'm not sure it has a valid foundation. Here's my thoughts ... * NSS provides a variety of utilities which utilize the secutil functions, certutil being the most prominent. Wouldn't the above argument mean you should remove certutil and friends from the distribution? I don't see that happening. * i18n by convention in the open source world is implemented in a manner which accommodates different implementations. Translatable strings are marked with _() and _() is defined in a (configurable) header file. To the best of my knowledge this works with all i18n implementations I'm aware of. GNU gettext is a standard in the open source world and to the best of my knowledge has been adopted by Mozilla. If that doesn't work in some environment then _() can be defined to be an identity function/macro and/or someone can provide the necessary patches for their unique non-standard implementation if they need it. It's open source after all. * Will you allow a "certain corporate management" to dictate your stagnation for 14 years when this is supposed to be an open source project independent of corporate management? * Will the potential for bug reports concerning lack of translations keep NSS from progressing for 14 years? Have there been a raft of bug reports against certutil etc. concerning it's lack of translations? * It's really not hard to add translation support, if push came to shove I could provide the patches, but even without translation support I don't see the justification for hiding the utility functions. Thoughts?
The various utility programs are classified as QA test programs, and as such are not required to be fully Internationalized and localized. NSS has not been stagnant. Products based on NSS provide their own UI, witness PSM for Mozilla, and Sun's server admin tools, to name just a few. The code in nss/cmd/lib can be thought of as exemplary, showing the basics to those who need to provide their own UI code.
Severity: normal → S3
Severity: S3 → S4
Status: UNCONFIRMED → RESOLVED
Closed: 3 years ago
Priority: -- → P5
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.