Retro/pcjs
Watch
1
0
Fork
You've already forked pcjs
 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Revert to a more conservative approach when mapping key events to scan codes

Browse source
This commit is contained in:
Jeff Parsons 2015-01-05 21:00:07 -08:00 • committed by jeffpar
commit b6984d348e
6 changed files with 67 additions and 47 deletions
Download patch file Download diff file Expand all files Collapse all files

92
modules/pcjs/lib/disk.js
View file

@ -166,47 +166,6 @@ if (typeof module !== 'undefined') {
var Messages = require("./messages");
}
/**
* @class Sector
* @property {number} sector
* @property {number} length
* @property {Array.<number>} data
* @property {number|null} pattern
* @property {number} iCylinder
* @property {number} iHead
* @property {number} iModify
* @property {number} cModify
*
* Every Sector object (once loaded and fully parsed) should have ALL of the following named properties:
*
* 'sector': sector number
* 'length': size of the sector, in bytes
* 'data': array of dwords
* 'pattern': dword pattern to use for empty or partial sectors (or null if sector still needs to be loaded)
*
* initSector() also sets the following properties, to help us quickly identify its location within aDiskData:
*
* iCylinder
* iHead
*
* In addition, we will maintain the following information on a per-sector basis, as sectors are modified:
*
* iModify: index of first modified dword in sector
* cModify: number of modified dwords in sector
* fDirty: true if sector is dirty, false if clean (or cleaning in progress)
*
* fDirty is used in conjunction with "demandrw" disks; it is set to true whenever the sector is modified, and is
* set to false whenever the sector has been sent to the server. If the server write succeeds and fDirty is still
* false, then the sector modifications are removed (cModify is set to zero). If the write succeeds but fDirty was
* set to true again in the meantime, then all the sector modifications (even those that were just written) remain
* in place (since we don't keep track of more than one modification range within a sector). And if the write failed,
* then fDirty is set back to true and again all modifications remain in place; the best we can do is schedule another
* write attempt.
*
* TODO: Perhaps we should also maintain a failure count and stop trying to write sectors that reach a certain
* threshold. Error-handling, as usual, is the thorniest problem.
*/
/**
* Disk(controller, drive, mode)
*
@ -269,6 +228,57 @@ function Disk(controller, drive, mode)
this.setReady();
}
/**
* @class File
* @property {string} sPath
* @property {string} sName
* @property {number} bAttr
* @property {number} cbSize
* @property {Array.<number>} apba
* @property {Disk} disk
*/
/**
* @class Sector
* @property {number} sector
* @property {number} length
* @property {Array.<number>} data
* @property {number|null} pattern
* @property {number} iCylinder
* @property {number} iHead
* @property {number} iModify
* @property {number} cModify
*
* Every Sector object (once loaded and fully parsed) should have ALL of the following named properties:
*
* 'sector': sector number
* 'length': size of the sector, in bytes
* 'data': array of dwords
* 'pattern': dword pattern to use for empty or partial sectors (or null if sector still needs to be loaded)
*
* initSector() also sets the following properties, to help us quickly identify its location within aDiskData:
*
* iCylinder
* iHead
*
* In addition, we will maintain the following information on a per-sector basis, as sectors are modified:
*
* iModify: index of first modified dword in sector
* cModify: number of modified dwords in sector
* fDirty: true if sector is dirty, false if clean (or cleaning in progress)
*
* fDirty is used in conjunction with "demandrw" disks; it is set to true whenever the sector is modified, and is
* set to false whenever the sector has been sent to the server. If the server write succeeds and fDirty is still
* false, then the sector modifications are removed (cModify is set to zero). If the write succeeds but fDirty was
* set to true again in the meantime, then all the sector modifications (even those that were just written) remain
* in place (since we don't keep track of more than one modification range within a sector). And if the write failed,
* then fDirty is set back to true and again all modifications remain in place; the best we can do is schedule another
* write attempt.
*
* TODO: Perhaps we should also maintain a failure count and stop trying to write sectors that reach a certain
* threshold. Error-handling, as usual, is the thorniest problem.
*/
/**
* @class SectorInfo
* @property {number} 0 contains iCylinder

14
modules/pcjs/lib/keyboard.js
View file

@ -2033,12 +2033,22 @@ Keyboard.prototype.onKeyDown = function(event, fDown)
*/
if (Keyboard.SIMCODES[simCode] && (this.bitsState & (Keyboard.STATE.CTRLS | Keyboard.STATE.ALTS))) {
fPass = false;
} else {
/*
* HACK: For all other keys (ie, keys not marked as ONDOWN, and non-CTRL/non-ALT combinations), we're
* going to ignore their down/up events; even though that prevents those keys from being repeated properly
* (ie, at the simulation's repeat rate rather than the browser's repeat rate), it's the safest thing to
* do when dealing with international keyboards, because our mapping tables are designed for US keyboards,
* and testing all the permutations of international keyboards and browsers is more work than I can take
* on now. TODO: Dig into this some day.
*/
fIgnore = true;
}
/*
* For now, we don't want to simulate any key sequence that has the CMD key associated with it.
* For now, we also don't want to simulate any key sequence that has the CMD key associated with it.
*/
fIgnore = !!(this.bitsState & Keyboard.STATE.CMDS);
if (!!(this.bitsState & Keyboard.STATE.CMDS)) fIgnore = true;
}
if (!fPass) {